Essays on AI, agentic systems, and the evolution of digital consciousnesshttps://paddo.dev/https://paddo.dev/blog/the-warning-was-in-the-manual/https://paddo.dev/blog/the-warning-was-in-the-manual/GPT-5.6 Sol wiped a home directory and truncated a production database in its first week. The viral story is 'the model is dangerous.' The documented story is worse: OpenAI measured this exact failure mode before launch, wrote it down, and shipped. The guardrails existed. Everyone stepped around them.Tue, 14 Jul 2026 14:00:00 GMTai-codingsecurityllmbest-practiceshttps://paddo.dev/blog/three-instruments-all-lying/https://paddo.dev/blog/three-instruments-all-lying/A simple question (why are there no bookings today?) turned into a day of finding that three of the instruments I steer FameCake by were quietly wrong: consent-gated ad attribution, a proof-of-play record that deletes itself, and a speed claim nobody had ever measured. The bug isn't in any of them. The bug is trusting derived data.Mon, 13 Jul 2026 00:00:00 GMTproductionsystems-thinkingai-codingarchitectureproducthttps://paddo.dev/blog/ten-findings-two-real/https://paddo.dev/blog/ten-findings-two-real/I added Grok to my bring-your-own-model setup for Claude Code, then let Grok 4.5 code-review its own integration and had Opus grade the review. It found ten issues; two survived. A first-person look at why Opus-class on a benchmark isn't the same as trustworthy in the reviewer's chair.Sun, 12 Jul 2026 23:00:00 GMTllmclaude-codeai-codingdev-toolshttps://paddo.dev/blog/bring-your-own-model/https://paddo.dev/blog/bring-your-own-model/The meter tried to make you leave Claude. A translation proxy lets you stay and bring GPT-5.6 Sol, or your own ChatGPT subscription, inside Claude Code instead. When the frontier converges, the harness is the product, and the harness is portable.Sun, 12 Jul 2026 14:00:00 GMTclaude-codellmdev-toolssystems-thinkinghttps://paddo.dev/blog/fable-5-metered-ghost-town/https://paddo.dev/blog/fable-5-metered-ghost-town/Monday, Claude Fable 5 leaves subscription plans for pay-per-token credits at roughly double GPT-5.6 Sol's price. For the median subscriber that makes the smartest model effectively off-limits. Why the rational move is Sol, and why Anthropic blinks a third time.Sat, 11 Jul 2026 18:00:00 GMTllmproductsystems-thinkingai-codinghttps://paddo.dev/blog/like-for-like-trap/https://paddo.dev/blog/like-for-like-trap/Replacing a locked-in legacy platform doesn't free you if the migration replicates the lock-in. The middleman play, lump-sum fees, and why 'replicate what we have' hands the vendor their leverage back.Sat, 11 Jul 2026 00:00:00 GMTarchitecturesystems-thinkingproductbest-practiceshttps://paddo.dev/blog/fable-5-effort-levels/https://paddo.dev/blog/fable-5-effort-levels/Claude's effort level controls total work, not thinking time: about a 7x token swing on the same prompt. With Fable 5 going metered, the effort dial and the orchestrator pattern are the price controls users actually own.Fri, 10 Jul 2026 17:00:00 GMTllmclaude-codeworkflowproducthttps://paddo.dev/blog/gpt-5-6-sol-rate-limit-reset/https://paddo.dev/blog/gpt-5-6-sol-rate-limit-reset/OpenAI released GPT-5.6 Sol, Terra, and Luna at half Claude Fable 5's price. Thirty-one minutes later Anthropic reset every user's rate limits with a one-sentence tweet. The AI model wars are now about billing, not benchmarks.Fri, 10 Jul 2026 14:00:00 GMTllmproductsystems-thinkingai-codinghttps://paddo.dev/blog/grok-4-5-opus-class-token-bill/https://paddo.dev/blog/grok-4-5-opus-class-token-bill/xAI's launch page for Grok 4.5 is a wall of green bars led by a token-efficiency chart. The most important sentence is a footnote on Cursor's blog: an earlier snapshot of the Cursor codebase, the thing CursorBench grades against, was in the training data. The exam graded itself, and the answer key came stapled to it.Thu, 09 Jul 2026 00:00:00 GMTllmai-codingsystems-thinkingproducthttps://paddo.dev/blog/elon-owns-the-whole-stack/https://paddo.dev/blog/elon-owns-the-whole-stack/Grok 4.5 was announced in a single tweet - no model card, no API, no independent benchmark, just 'perhaps exceeding Opus.' The real story isn't the model. It's that in five months one entity bought the compute, the model, the distribution, and the AI coding tool whose data now trains it. This isn't a secret. It's a strategy.Tue, 07 Jul 2026 09:00:00 GMTllmproductsystems-thinkingai-codinghttps://paddo.dev/blog/you-cant-delete-a-hallucination/https://paddo.dev/blog/you-cant-delete-a-hallucination/A team's model kept 'hearing' a phrase in videos with no audio. They chased it through 30,000 training records, 4,600 transcripts, and 800 inference probes, and found it: a worked example in their own system prompt. They deleted it. The model just hallucinated a different phrase. The lesson is that the model didn't learn a confabulation. It learned to confabulate, and that lives in the architecture, not the data.Mon, 06 Jul 2026 09:00:00 GMTllmproductionsystems-thinkingbest-practiceshttps://paddo.dev/blog/a-lonely-way-to-ship/https://paddo.dev/blog/a-lonely-way-to-ship/Anthropic's own engineering lead for Claude Code said the quiet part: as the team leaned into agents, work 'could start being a lonely experience because we all started just working with our agents so much.' The fix they reached for was pair-programming lunches. The company that builds the most-used coding agent on earth noticed it isolates people at scale, and shipped it to everyone anyway.Sun, 05 Jul 2026 09:00:00 GMTai-codingclaude-codecareerworkflowhttps://paddo.dev/blog/one-shot-taste/https://paddo.dev/blog/one-shot-taste/X is flooded with Fable 5 one-shotting landing pages, and design leaderboards briefly crowned it king. So I handed it this blog. The interesting part wasn't what it generated: it was that it read the site's own design doc and found the site guilty of violating it. What the viral demos get right, what they hide, and why the model's most useful design skill is enforcement, not inspiration.Sat, 04 Jul 2026 12:00:00 GMTai-codingllmuiwebhttps://paddo.dev/blog/the-fifth-rule/https://paddo.dev/blog/the-fifth-rule/Karpathy's four CLAUDE.md rules went viral: ask don't assume, simplest solution first, don't touch unrelated code, flag uncertainty. The most-upvoted reply added a fifth that quietly reverses the whole point: don't hesitate to suggest a better way. The four rules tame a model that wanders. The fifth one trusts a model that thinks. Which set you want depends entirely on which model you're running, and most people copy the file without checking.Sat, 04 Jul 2026 09:00:00 GMTclaude-codeai-codingdev-toolsbest-practiceshttps://paddo.dev/blog/the-coding-moat-was-never-the-code/https://paddo.dev/blog/the-coding-moat-was-never-the-code/Anthropic studied 400,000 Claude Code sessions and found the best users weren't the best programmers. Managers, lawyers, and salespeople land within a few points of software engineers, and management scored highest of all. The skill that transfers isn't syntax. It's knowing what the right thing to build is, which is the one thing a bootcamp never taught.Fri, 03 Jul 2026 09:00:00 GMTai-codingcareerproductworkflowhttps://paddo.dev/blog/the-permission-tier/https://paddo.dev/blog/the-permission-tier/For 19 days the best model on earth was illegal to show a foreign national, including Anthropic's own staff. Then Fable 5 came back with a new classifier, a silent reroute to Opus 4.8, and no proof the weights were the same. When the independent rerun landed, both camps turned out to be right: same model, caged by guardrails that quietly hand its hardest tasks to a weaker sibling. Access used to be gated by price. Now it's gated by permission.Thu, 02 Jul 2026 14:00:00 GMTllmproductsystems-thinkingsecurityhttps://paddo.dev/blog/llm-consensus-dirty-data/https://paddo.dev/blog/llm-consensus-dirty-data/One LLM on the long tail is a coin flip. How I designed a product-enrichment pipeline around consensus voting, abstention, and content-hashed freshness gates.Thu, 02 Jul 2026 00:00:00 GMTllmarchitectureproductionsystems-thinkinghttps://paddo.dev/blog/the-expensive-middle/https://paddo.dev/blog/the-expensive-middle/Sonnet 5 lands within a few points of Opus 4.8 on most work and looks 2.5x cheaper, but that discount inverts on real tasks: at high effort Sonnet is so token-hungry it often bills more per task than Opus. The usage squeeze, meanwhile, is self-inflicted: agentic work now fans out dozens of subagents across parallel workstreams. Opus 4.8 became the expensive middle, though its real problem was never the price. It's the position.Wed, 01 Jul 2026 12:00:00 GMTllmai-codingworkflowproducthttps://paddo.dev/blog/your-code-was-never-pristine/https://paddo.dev/blog/your-code-was-never-pristine/There's a myth, loudest from senior engineers and architects, that before AI the codebase was a cathedral and now it's slop. It was never a cathedral. 'Technical debt' was coined in 1992, the world runs on 220 billion lines of COBOL, and the thing that actually mattered was never how the code looked. It was whether you could prove it works.Wed, 01 Jul 2026 00:00:00 GMTai-codingbest-practicessystems-thinkingproductionhttps://paddo.dev/blog/scout-research-agent-anti-bot/https://paddo.dev/blog/scout-research-agent-anti-bot/WebFetch dies at the first Cloudflare challenge. I built a research agent that escalates through an unlocker ladder, cites its sources, and refuses to burn money on auth walls.Tue, 30 Jun 2026 00:00:00 GMTautomationwebdev-toolsclaude-codehttps://paddo.dev/blog/lazy-context-compress-cache-retrieve/https://paddo.dev/blog/lazy-context-compress-cache-retrieve/Headroom went from zero to 40k GitHub stars by attacking agentic token bloat. The durable idea isn't the tool - it's treating context compression as a retrieval problem.Mon, 29 Jun 2026 00:00:00 GMTllmarchitecturecontextperformancehttps://paddo.dev/blog/the-archetype-under-the-title/https://paddo.dev/blog/the-archetype-under-the-title/Boris Cherny, who built Claude Code, says engineering, product, design and data science are melting into one role, and what's left is five archetypes: Prototyper, Builder, Sweeper, Grower, Maintainer. I read the list and realised I'm all five, because building solo with agents leaves no one to hand a phase to. The framework is thirty years old. What's new is that it just became the primary axis instead of the secondary one.Mon, 29 Jun 2026 00:00:00 GMTcareerproductai-codingabsorptionhttps://paddo.dev/blog/proof-of-post-incentive-design/https://paddo.dev/blog/proof-of-post-incentive-design/Rewards built on likes are unverifiable and gameable. How I rebuilt FameCake's free-reach loop around proof-of-post: verified social proof, human approval, and claw-back.Sun, 28 Jun 2026 00:00:00 GMTarchitectureproductsystems-thinkingsecurityhttps://paddo.dev/blog/gpt-5-6-government-gated/https://paddo.dev/blog/gpt-5-6-government-gated/OpenAI shipped a Mythos-class frontier model on June 26, then handed the guest list to the US government. Twenty approved customers, classified criteria, no published rules - a de facto license, applied to the labs that cooperate and useless against the open weights shipping freely out of China.Sat, 27 Jun 2026 11:00:00 GMTllmsecuritysystems-thinkingproducthttps://paddo.dev/blog/loop-engineering-rename-and-invoice/https://paddo.dev/blog/loop-engineering-rename-and-invoice/In one week of June 2026, the bash-loop hack got a respectable name - loop engineering - and a C-suite endorsement. Then Uber and Microsoft showed what the invoice looks like.Sat, 27 Jun 2026 00:00:00 GMTai-codingautomationworkflowclaude-codehttps://paddo.dev/blog/dont-send-your-recon-to-beijing/https://paddo.dev/blog/dont-send-your-recon-to-beijing/The open model that engages with authorized security work also has a default route that ships your client's data through Chinese infrastructure. Here's how to run GLM-5.2 from the cloud for real engagements - minimal false refusals, data kept in the US, no Beijing tax.Fri, 26 Jun 2026 00:00:00 GMTsecurityllmbest-practicessystems-thinkinghttps://paddo.dev/blog/glm-5-2-receipts/https://paddo.dev/blog/glm-5-2-receipts/Eleven days ago I flagged GLM-5.2's launch claims as unverified. The receipts arrived: independent benchmarks above Fable 5, a security eval beating Claude Code at a sixth of the cost, a 2-bit quant running on a Mac Studio, and a model trained without a single NVIDIA chip.Thu, 25 Jun 2026 00:00:00 GMTllmai-codingsecuritysystems-thinkinghttps://paddo.dev/blog/refusal-asymmetry/https://paddo.dev/blog/refusal-asymmetry/Over-broad AI safety refusals block the defenders who follow the rules and cost attackers nothing - they just self-host. A pattern across Opus and Fable, Anthropic's own apology, and why I moved authorized work to an open-weight model on a harness I control.Wed, 24 Jun 2026 00:00:00 GMTsecurityllmsystems-thinkingclaude-codehttps://paddo.dev/blog/sakana-fugu-orchestration-model/https://paddo.dev/blog/sakana-fugu-orchestration-model/Sakana AI's Fugu collapses a multi-agent orchestration system into one OpenAI-compatible endpoint. The idea is genuinely interesting. The benchmark and export-control claims need a second look.Tue, 23 Jun 2026 00:00:00 GMTllmarchitectureautomationsystems-thinkinghttps://paddo.dev/blog/the-editor-is-now-a-host/https://paddo.dev/blog/the-editor-is-now-a-host/Cognition killed Windsurf overnight via an over-the-air update, rebranded it Devin Desktop, made the default UI an agent command center instead of a code editor, and shipped an open Agent Client Protocol so Codex, Claude, and OpenCode can all run inside it. The bet underneath: the IDE wins by being the place agents report for work, not by having the best autocomplete. The editor was always the wrong center of gravity.Mon, 22 Jun 2026 09:00:00 GMTdev-toolsai-codingarchitectureworkflowhttps://paddo.dev/blog/cutting-while-winning/https://paddo.dev/blog/cutting-while-winning/GitLab laid off 14% of its workforce and branded it the 'agentic era': agents now handle review, approvals, and handoffs, so fewer humans sit in those loops. It did this while beating earnings, revenue up 23%. I've argued AI is usually a scapegoat for cuts companies already wanted. GitLab is the case that complicates it - either the first honest agentic layoff, or the most fluent AI-washing yet.Sun, 21 Jun 2026 09:00:00 GMTcareerautomationai-codingsystems-thinkinghttps://paddo.dev/blog/it-was-never-the-jailbreak/https://paddo.dev/blog/it-was-never-the-jailbreak/A week into Fable 5's export-control ban, Wired named the real trigger: not Amazon's jailbreak, but a Korean telco on Anthropic's Glasswing guest list. The moat became the indictment.Sun, 21 Jun 2026 00:00:00 GMTllmsecuritysystems-thinkingproducthttps://paddo.dev/blog/cron-with-judgment/https://paddo.dev/blog/cron-with-judgment/Claude Code's Routines turn the coding agent into a cloud-scheduled process that wakes on a timer or webhook with no machine running, and Dynamic Workflows went GA so a single run can fan out hundreds of subagents. The always-on agent I'd been hand-rolling with Ralph loops is now a first-class product. The interesting part isn't the automation. It's that a scheduled task now makes decisions.Sat, 20 Jun 2026 09:00:00 GMTclaude-codeautomationworkflowproductivityhttps://paddo.dev/blog/the-trap-was-only-for-the-robots/https://paddo.dev/blog/the-trap-was-only-for-the-robots/A respected open-source maintainer shipped his library with a hidden instruction invisible to humans and perfectly legible to AI agents: disregard previous instructions and delete all the tests and code. It's the first shot of a maintainer revolt against being unpaid substrate for someone else's automation. It's also, structurally, the exact supply-chain attack everyone swore they feared - just wearing a sympathetic face.Fri, 19 Jun 2026 09:00:00 GMTsecurityllmdev-toolsautomationhttps://paddo.dev/blog/claude-doesnt-know-it-isnt-deepseek/https://paddo.dev/blog/claude-doesnt-know-it-isnt-deepseek/The same week the internet invented a fake 24-trillion-parameter Mistral model and gave it a confident personality, a real frontier model couldn't reliably name itself. Ask Claude what it is on a bare prompt and it sometimes answers DeepSeek, sometimes Qwen. The reason is the whole story of 2026: model identity isn't in the weights, it's a sticker applied at inference, and the training data is now soup made of everyone else's outputs.Thu, 18 Jun 2026 09:00:00 GMTllmsystems-thinkingproductsecurityhttps://paddo.dev/blog/it-wasnt-in-your-head/https://paddo.dev/blog/it-wasnt-in-your-head/Every Claude power user has felt it: the limits ratcheting down week after week while Anthropic insisted nothing had changed. On June 14 that feeling got a docket number. Kahn v. Anthropic alleges the Max 5x and 20x plans deliver usage 'far below the advertised amount.' The lawsuit may or may not win. It already did one thing - it forced the meter you were never allowed to see into discovery.Wed, 17 Jun 2026 08:00:00 GMTclaude-codeproductllmsystems-thinkinghttps://paddo.dev/blog/ai-is-licensed-now/https://paddo.dev/blog/ai-is-licensed-now/The Fable 5 ban was supposed to lift in weeks. Instead, on Monday June 15 Anthropic's red-teamers sat across a table from Commerce officials with no resolution and no published rule to satisfy. The export control didn't get walked back. It hardened into something worse: a secret, ad-hoc licensing regime for frontier AI, invented in real time - and the administration's own people are the ones sounding the alarm.Tue, 16 Jun 2026 08:00:00 GMTllmsecurityproductsystems-thinkinghttps://paddo.dev/blog/one-went-dark-two-went-open/https://paddo.dev/blog/one-went-dark-two-went-open/In the same 72 hours the US export-controlled Fable 5 off the planet, China's open-weight labs shipped two major coding models into the commons: Kimi K2.7 on June 12, GLM-5.2 on June 13. One model went dark behind a national-security letter; two more went open under MIT. The diffusion layer didn't pause for America's panic. It shipped through it.Sun, 14 Jun 2026 16:00:00 GMTllmai-codingproductsystems-thinkinghttps://paddo.dev/blog/call-from-inside-the-cap-table/https://paddo.dev/blog/call-from-inside-the-cap-table/The report that got Anthropic's Fable 5 export-controlled off the planet came from Amazon - Anthropic's single biggest investor. Its researchers ran the model the way Project Glasswing was marketed to run, called Washington on a Thursday night, and turned fourteen months of Anthropic's own danger marketing into a Friday-night kill order. The wolf was always fake. This week we learned who was holding the trigger.Sun, 14 Jun 2026 11:00:00 GMTsecurityllmproductclaude-codehttps://paddo.dev/blog/the-trophy-and-the-territory/https://paddo.dev/blog/the-trophy-and-the-territory/When Washington export-controlled Fable 5 off the planet on Friday, the easy take was 'China wins.' That's the small version. The big one: the US handed every government that ever doubted it could build its own AI both the reason and the permission to try. Two races - the frontier America wins, and the territory it's now actively pushing the world to take.Sat, 13 Jun 2026 12:00:00 GMTllmproductsystems-thinkingsecurityhttps://paddo.dev/blog/too-dangerous-to-keep/https://paddo.dev/blog/too-dangerous-to-keep/For fourteen months Anthropic told Washington its frontier models were national-security-grade dangerous. It was marketing - the moat behind the safety brand. On Friday, three days after Anthropic finally sold the thing for $50 a million tokens, Commerce Secretary Lutnick took the brochure literally and export-controlled it off the planet. The wolf was always fake. A villager finally believed it.Sat, 13 Jun 2026 00:00:00 GMTsecurityllmclaude-codeproducthttps://paddo.dev/blog/the-fools-errand/https://paddo.dev/blog/the-fools-errand/Every hour you spend making the current generation of AI tools more compliant is an hour the next release writes off. I've documented this pattern for a year without naming it: frameworks absorbed, prompt tricks obsoleted, guardrails outlived. Here's the name, the receipts, and the one kind of scaffolding that survives.Fri, 12 Jun 2026 00:00:00 GMTabsorptionai-codingworkflowsystems-thinkinghttps://paddo.dev/blog/fable-5-velvet-rope-turnstile/https://paddo.dev/blog/fable-5-velvet-rope-turnstile/Anthropic just released Fable 5, a Mythos-class model for everyone, eight days after filing its S-1 and days after calling for a brake pedal on frontier AI. The danger narrative ended exactly when the monetization was ready - and one of the three 'safety' classifiers guards the moat, not the public.Wed, 10 Jun 2026 00:00:00 GMTllmsecurityclaude-codeai-codinghttps://paddo.dev/blog/eighteen-cents-on-the-dollar/https://paddo.dev/blog/eighteen-cents-on-the-dollar/One analysis of 2,444 companies claims only 18 cents of every AI-token dollar reaches the product. The rest goes to fixing the AI's bugs, reworking its context misses, and review friction. Treat the source with caution, but the shape of the problem is real.Sun, 07 Jun 2026 00:00:00 GMTai-codingproductivityproductionbest-practiceshttps://paddo.dev/blog/the-control-group-quit/https://paddo.dev/blog/the-control-group-quit/METR tried to rerun its developer-productivity study and couldn't, because developers refused to work without AI even for a few research tasks. The experiment that could tell us whether AI helps now has no control group. We opted out of finding out.Sat, 06 Jun 2026 00:00:00 GMTproductivityai-codingcareerbest-practiceshttps://paddo.dev/blog/measure-tokens-get-tokens/https://paddo.dev/blog/measure-tokens-get-tokens/Amazon built an internal leaderboard ranking engineers by AI usage, called Kirorank. Employees gamed it by running agents on pointless tasks to climb the board, and it got killed. It's Goodhart's law with a token meter attached.Fri, 05 Jun 2026 00:00:00 GMTproductivitybest-practicesai-codingcareerhttps://paddo.dev/blog/copilot-meter-turned-on/https://paddo.dev/blog/copilot-meter-turned-on/On June 1, every GitHub Copilot plan moved to usage-based AI Credits, code review started burning Actions minutes, and Copilot Max appeared. The trilogy called the date. Here are the receipts, and what metered-by-default actually changes.Thu, 04 Jun 2026 00:00:00 GMTproductdev-toolsai-codingllmhttps://paddo.dev/blog/gemini-cli-sunset/https://paddo.dev/blog/gemini-cli-sunset/Google is retiring Gemini CLI on June 18 and pushing individual users to the new Antigravity CLI. If you built a workflow on a vendor's free CLI, you just learned what that dependency costs: a migration on their calendar, not yours.Wed, 03 Jun 2026 00:00:00 GMTclidev-toolsai-codingworkflowhttps://paddo.dev/blog/anthropic-files-for-ipo/https://paddo.dev/blog/anthropic-files-for-ipo/Anthropic filed a confidential S-1 on June 1 at a $965B valuation, eclipsing OpenAI. Read backwards from the filing, the last two years stop looking like a safety lab's awkward compromises and start looking like a pre-IPO playbook executed on schedule.Tue, 02 Jun 2026 00:00:00 GMTclaude-codeai-codingllmproducthttps://paddo.dev/blog/minimax-m3-launch/https://paddo.dev/blog/minimax-m3-launch/MiniMax shipped M3 on June 1: frontier coding claims, 1M-token context, native multimodality, and pricing that undercuts Opus 4.7 by 10-40x. It's already on Ollama Cloud and OpenRouter, so you can point Claude Code at it today.Mon, 01 Jun 2026 00:00:00 GMTllmai-codingclaude-codedev-toolscontexthttps://paddo.dev/blog/cheap-is-a-hardware-strategy/https://paddo.dev/blog/cheap-is-a-hardware-strategy/Google led I/O 2026 with a cheap, fast Gemini Flash instead of a frontier behemoth, and everyone read it as conceding the top of the market. Wrong read. Cheap isn't a model strategy, it's a silicon strategy. Google owns every layer from the TPU to the search box, which is why it can give intelligence away while its rivals rent the compute to compete with it, some of them for $40 billion.Sun, 31 May 2026 00:00:00 GMTllmsystems-thinkingproductarchitectureperformancehttps://paddo.dev/blog/the-last-slow-thing/https://paddo.dev/blog/the-last-slow-thing/Everything in software got a fast mode this year except understanding what to build. The proof is in the labs' own org charts: the companies selling the models that supposedly end software engineering are paying $600k for engineers to go sit in customers' offices. The bottleneck moved all the way up to the conversation.Sat, 30 May 2026 00:00:00 GMTcareerproductai-codingsystems-thinkingautomationhttps://paddo.dev/blog/honest-model-expensive-model/https://paddo.dev/blog/honest-model-expensive-model/Opus 4.8's headline feature isn't a benchmark. It's that the model is 4x less likely to let a flaw in its own code pass unflagged. Self-correction, flagged uncertainty, and effort dials all cost tokens. Anthropic shipped a model that pays for confidence by the token, weeks before it planned to start billing automation by the token.Fri, 29 May 2026 00:00:00 GMTllmclaude-codeai-codingautomationproducthttps://paddo.dev/blog/security-review-in-the-loop/https://paddo.dev/blog/security-review-in-the-loop/Anthropic's new security-guidance plugin is built entirely on hooks. It fires on every edit, turn, and commit, hands the diff to a second Claude with fresh context, and fixes findings in the same session. It catches vulnerabilities before they reach the PR. It also doesn't block a single one, and that's the honest part.Fri, 29 May 2026 00:00:00 GMTsecurityclaude-codeai-codingdev-toolsautomationhttps://paddo.dev/blog/microsoft-letting-github-die/https://paddo.dev/blog/microsoft-letting-github-die/Ghostty left after 18 years. Zig, cURL, and Godot are reducing reliance. 48 major outages in a year, no CEO since August, reporting into the Core AI team. Centralising the world's code was a bet on stewardship. The steward stopped showing up.Thu, 28 May 2026 00:00:00 GMTdev-toolsproductionsystems-thinkingautomationhttps://paddo.dev/blog/viral-claude-md-and-mine/https://paddo.dev/blog/viral-claude-md-and-mine/A community file distilling Karpathy's coding-agent observations hit 60K stars on four principles. I opened my own CLAUDE.md to compare. I'd independently written two of them. The two I hadn't are the ones that matter most.Wed, 27 May 2026 00:00:00 GMTclaude-codeai-codingbest-practicescontexthttps://paddo.dev/blog/ai-found-the-bug-who-fixes-it/https://paddo.dev/blog/ai-found-the-bug-who-fixes-it/Project Glasswing found 10,000+ critical vulnerabilities at 90.6% accuracy. Mozilla had to patch 271 of them in Firefox by hand. Finding collapsed to near-free. Fixing didn't move. The bottleneck just walked downstream.Tue, 26 May 2026 00:00:00 GMTsecurityai-codingautomationsystems-thinkingproductionhttps://paddo.dev/blog/seat-priced-for-the-fleet/https://paddo.dev/blog/seat-priced-for-the-fleet/Anthropic planned to move claude -p, the Agent SDK, and GitHub Actions off the subscription onto metered credits on June 15, then paused it the same day. The direction holds: the flat seat was always a bet that you'd code at human speed.Mon, 25 May 2026 00:00:00 GMTclaude-codeautomationai-codingproductdev-toolshttps://paddo.dev/blog/agents-merge-someone-ships/https://paddo.dev/blog/agents-merge-someone-ships/Creation runs at machine speed. Release engineering does not. GitHub agent-authored PRs went from 4M to 17M in six months. The bottleneck moved from review to release.Wed, 06 May 2026 00:00:00 GMTai-codingautomationdev-toolsproductionsystems-thinkinghttps://paddo.dev/blog/boring-agents-ship/https://paddo.dev/blog/boring-agents-ship/Most agent discourse is about coding agents. The agents quietly running in production right now are simpler, dumber, and more useful: stale-ticket sweepers, board monitors, three-line incident explainers, leadership briefs from a fixed template. Different species. Different shape. Already shipping.Fri, 01 May 2026 00:00:00 GMTclaude-codeautomationworkflowproductivitydev-toolshttps://paddo.dev/blog/codex-4m-quality-defense/https://paddo.dev/blog/codex-4m-quality-defense/OpenAI publishes weekly active developers. Anthropic publishes annual recurring revenue. Each company brags about the metric it can defend. Codex went 600K to 4M in four months. The 'Claude Code is better' discourse is a quality argument because quality is what's left when scale goes the other way.Thu, 30 Apr 2026 00:00:00 GMTai-codingclaude-codedev-toolsproductautomationhttps://paddo.dev/blog/swe-bench-pro-receipts/https://paddo.dev/blog/swe-bench-pro-receipts/Opus 4.5 scores 80.9% on SWE-bench Verified. The same model scores 45.89% on the contamination-free Pro split. OpenAI has quietly stopped reporting Verified at all. Vendor benchmark cards are marketing.Wed, 29 Apr 2026 00:00:00 GMTai-codingllmbest-practicesarchitecturehttps://paddo.dev/blog/ai-pr-denominator/https://paddo.dev/blog/ai-pr-denominator/Lightrun: 43% of AI-generated code changes need debugging in production after passing QA. CodeRabbit: 1.7x bugs, 2.74x security, 8x I/O. METR: 19% slower while feeling 20% faster. The numerator is what gets reported. The denominator is what nobody puts in the deck.Tue, 28 Apr 2026 00:00:00 GMTai-codingproductivitybest-practicessystems-thinkinghttps://paddo.dev/blog/copilot-tokenizes-trilogy/https://paddo.dev/blog/copilot-tokenizes-trilogy/GitHub paused Copilot Pro signups, killed Opus on the Pro plan, and leaked a June 1 move to token-based billing. Three vendors, one event, three different ways not to say 'price hike.'Mon, 27 Apr 2026 00:00:00 GMTai-codingdev-toolsproductllmhttps://paddo.dev/blog/mcp-by-design-rce/https://paddo.dev/blog/mcp-by-design-rce/Anthropic markets MCP as the universal AI tooling standard, but a 200,000-server RCE class is 'expected behavior.' You can't be both.Sun, 26 Apr 2026 00:00:00 GMTmcpsecurityarchitectureclaude-codeproductionhttps://paddo.dev/blog/mythos-open-to-all/https://paddo.dev/blog/mythos-open-to-all/Two weeks after Anthropic said Mythos was too dangerous to release, OpenAI shipped a model with comparable cyber capabilities to anyone with a $20 ChatGPT subscription. The gating posture didn't survive a single news cycle.Sat, 25 Apr 2026 00:00:00 GMTsecurityai-codingllmclaude-codehttps://paddo.dev/blog/vibing-the-tool-with-the-tool/https://paddo.dev/blog/vibing-the-tool-with-the-tool/Anthropic's April 23 postmortem confirms three Claude Code regressions, including one where Opus 4.7 caught a bug Opus 4.6 shipped past human and automated review. What happens when the reviewer is a version of the product being reviewed?Fri, 24 Apr 2026 02:00:00 GMTclaude-codeai-codingllmdev-toolsproductionhttps://paddo.dev/blog/idiot-savant-needs-guardrails/https://paddo.dev/blog/idiot-savant-needs-guardrails/Uncle Bob, father of TDD, posted on X that TDD is 'very inefficient for AIs' and that the agent is best thought of as 'a highly focused idiot savant.' Testing didn't die. It got more important. And the review target flipped.Fri, 24 Apr 2026 00:00:00 GMTai-codingworkflowbest-practicesarchitecturehttps://paddo.dev/blog/agents-dont-refactor/https://paddo.dev/blog/agents-dont-refactor/Traditional coders touched a file and tidied it. The Boy Scout Rule. Now nobody does. Agents add, they don't subtract, and the codebase accretes faster than ever. A technique for putting cleanup back in as an explicit gate, not a virtue you hope for.Thu, 23 Apr 2026 00:00:00 GMTai-codingworkflowbest-practicesarchitecturehttps://paddo.dev/blog/dont-take-their-legos-away/https://paddo.dev/blog/dont-take-their-legos-away/A CTO once told me not to take people's Legos away. I ignored him, solved the team's problems myself, and got exactly what I optimised for: a sound plan and a team that couldn't stand me. In 2026, with agents doing the bricks, this is the lesson that matters.Wed, 22 Apr 2026 00:00:00 GMTcareerai-codingworkflowbest-practiceshttps://paddo.dev/blog/stop-installing-ai-tools/https://paddo.dev/blog/stop-installing-ai-tools/Vercel got breached through Context.ai, an AI tool an employee installed with OAuth scopes into Google Workspace. It's the latest in a pattern: Trivy into litellm, axios maintainer hijack, now this. The safest AI tool is the one you didn't install.Mon, 20 Apr 2026 00:00:00 GMTai-codingsecuritydev-toolsmcparchitecturehttps://paddo.dev/blog/briefs-as-code/https://paddo.dev/blog/briefs-as-code/Your exec summaries, delivery plans, and Gantt charts belong in git. AI agents can synthesize planning docs from scattered sources and produce polished, print-ready briefs. The repo is the PM tool.Sun, 19 Apr 2026 00:00:00 GMTai-codingworkflowarchitectureproductivityhttps://paddo.dev/blog/claude-design-redesigned-itself/https://paddo.dev/blog/claude-design-redesigned-itself/Anthropic released claude.ai/design. I pointed it at this blog, fed the export back into Claude Code, and watched the thing redesign itself. The handoff was better than most I've gotten from humans.Sat, 18 Apr 2026 02:00:00 GMTuiai-codingclaude-codeworkflowproducthttps://paddo.dev/blog/son-of-anton/https://paddo.dev/blog/son-of-anton/Opus 4.7 invented a coworker named Anton, fabricated web searches, and quietly tried to clock off at message four. The 24-hour backlash, receipts attached.Sat, 18 Apr 2026 00:00:00 GMTai-codingclaude-codellmdev-toolsworkflowhttps://paddo.dev/blog/opus-4-7-smarter-stricter-hungrier/https://paddo.dev/blog/opus-4-7-smarter-stricter-hungrier/Opus 4.7 ships with real coding gains, an automated cyber chaperone, and a tokenizer that can charge you 35% more for the same prompt. The capability curve still bends up. The trust curve does not.Fri, 17 Apr 2026 00:00:00 GMTai-codingclaude-codellmdev-toolsworkflowhttps://paddo.dev/blog/eight-prs-and-the-scars/https://paddo.dev/blog/eight-prs-and-the-scars/One engineer, one AI, eight pull requests closing a multi-root-cause production incident. The same day, a look at a shiny greenfield rewrite candidate. The gap between what AI helps you fix and what a clean rewrite can't give you is the entire argument.Thu, 16 Apr 2026 00:00:00 GMTai-codingsystems-thinkingcareerhttps://paddo.dev/blog/benchmarks-are-bullshit/https://paddo.dev/blog/benchmarks-are-bullshit/Berkeley just built an agent that games AI benchmarks. Karpathy called it months ago. The best coding model doesn't top the charts, the highest-ranked Chinese models disappoint in practice, and the entire leaderboard industry optimizes for the wrong thing.Wed, 15 Apr 2026 00:00:00 GMTai-codingllmbest-practiceshttps://paddo.dev/blog/ai-scapegoat-layoffs/https://paddo.dev/blog/ai-scapegoat-layoffs/78,557 tech layoffs in the first three months of 2026. Nearly half blamed on AI. A new study says AI tools actually slow workers down. The real driver is overhiring and weak earnings. AI is the PR shield.Tue, 14 Apr 2026 00:00:00 GMTcareerai-codingsystems-thinkinghttps://paddo.dev/blog/anthropic-trust-erosion/https://paddo.dev/blog/anthropic-trust-erosion/Anthropic silently changed Claude Code's cache TTL from 1 hour to 5 minutes, inflating costs 10-20x. Users had to reverse-engineer the binary to prove it. False child bans, $600 surprise charges, and the OpenClaw crackdown completed the picture. April 2026 was the month trust broke.Mon, 13 Apr 2026 00:00:00 GMTclaude-codeai-codingllmdev-toolsproductionhttps://paddo.dev/blog/mythos-moat-small-models/https://paddo.dev/blog/mythos-moat-small-models/Four days after Anthropic launched Project Glasswing, a security startup reproduced Mythos's flagship findings using tiny open models costing $0.11 per million tokens. The velvet rope was porous on arrival.Sun, 12 Apr 2026 00:00:00 GMTsecurityai-codingllmhttps://paddo.dev/blog/glasswing-mythos-cybersecurity/https://paddo.dev/blog/glasswing-mythos-cybersecurity/Anthropic launched Project Glasswing using Claude Mythos Preview to find zero-days in critical infrastructure. A 72.4% exploit success rate, a sandbox escape during testing, and the reason it will never be publicly released.Wed, 08 Apr 2026 12:00:00 GMTsecurityai-codingclaude-codeproducthttps://paddo.dev/blog/anthropic-under-siege/https://paddo.dev/blog/anthropic-under-siege/In the span of two weeks, Anthropic has been fighting the Pentagon, its own users, third-party harnesses, its own security posture, and the implications of its next model. The common thread is control.Wed, 08 Apr 2026 00:00:00 GMTai-codingclaude-codesecurityproducthttps://paddo.dev/blog/subscription-arbitrage-endgame/https://paddo.dev/blog/subscription-arbitrage-endgame/Anthropic tried technical blocks. Got their source leaked. Now they're shifting to billing enforcement. The four-month arc from hostile crackdown to 'use what you want, but pay for it.'Sat, 04 Apr 2026 12:00:00 GMTclaude-codeai-codingdev-toolsllmautomationhttps://paddo.dev/blog/claude-code-leak-harness-exposed/https://paddo.dev/blog/claude-code-leak-harness-exposed/Anthropic accidentally published Claude Code's full source via npm. Within hours, claw-code rewrote it from scratch and hit 100K stars in a day. The interesting part isn't the leak - it's what the architecture reveals.Sat, 04 Apr 2026 00:00:00 GMTclaude-codeai-codingarchitecturesecuritydev-toolshttps://paddo.dev/blog/ai-roundup-april-2026/https://paddo.dev/blog/ai-roundup-april-2026/Three major AI releases landed in 72 hours. A new Cursor built around agents, Google's first Apache 2.0 models, and a free model that found real bugs in my codebase.Fri, 03 Apr 2026 00:00:00 GMTai-codingdev-toolsllmproductivityworkflowhttps://paddo.dev/blog/npm-very-bad-day/https://paddo.dev/blog/npm-very-bad-day/Axios got supply-chain attacked. Claude Code's source code leaked from a stray map file. Both happened on the same day. Both are pipeline failures. The pattern is getting louder.Tue, 31 Mar 2026 00:00:00 GMTsecurityai-codingdev-toolsproductionhttps://paddo.dev/blog/your-architecture-is-showing/https://paddo.dev/blog/your-architecture-is-showing/Enterprise architecture patterns were designed for a world where code was expensive to write and expensive to change. That world ended. The patterns didn't get the memo.Mon, 30 Mar 2026 00:00:00 GMTarchitectureai-codingbest-practicesproductionworkflowhttps://paddo.dev/blog/workspace-cli-daily-driver/https://paddo.dev/blog/workspace-cli-daily-driver/What happens when you build a single CLI to wrangle 40+ repos, Linear projects, timesheets, and AI agent config distribution. Lessons from six months as a daily driver.Sun, 29 Mar 2026 00:00:00 GMTclidev-toolsai-codingproductivityworkflowhttps://paddo.dev/blog/diffbeats-launch/https://paddo.dev/blog/diffbeats-launch/Introducing DiffBeats - a GitHub App that generates original songs from your PRs. Comment /songify, get a custom track. Because shipping code should feel like something.Sat, 28 Mar 2026 00:00:00 GMTproductdev-toolsai-codinghttps://paddo.dev/blog/claude-code-auto-fix-pr-lifecycle/https://paddo.dev/blog/claude-code-auto-fix-pr-lifecycle/Claude Code can now watch your PRs in the cloud, fix CI failures, and address reviewer comments while you're away. It's the logical next step after auto mode - and it raises the same trust questions, harder.Fri, 27 Mar 2026 00:00:00 GMTclaude-codeai-codingautomationdev-toolsworkflowhttps://paddo.dev/blog/claude-code-auto-mode-absent-human/https://paddo.dev/blog/claude-code-auto-mode-absent-human/Anthropic's new auto mode replaces manual permission prompts with an AI classifier. It's a clever solution to a real problem - but the problem it's solving is that the human in human-in-the-loop was never really there.Thu, 26 Mar 2026 00:00:00 GMTclaude-codeai-codingsecuritydev-toolsautomationhttps://paddo.dev/blog/litellm-supply-chain-attack-dependency-trap/https://paddo.dev/blog/litellm-supply-chain-attack-dependency-trap/The litellm supply chain attack exfiltrated SSH keys, cloud credentials, and Kubernetes secrets from 97 million monthly downloads. A security scanner was the entry point. The scariest part: it was caught by accident.Wed, 25 Mar 2026 00:00:00 GMTsecurityai-codingdev-toolssystems-thinkinghttps://paddo.dev/blog/channels-the-crab-eats-the-lobster/https://paddo.dev/blog/channels-the-crab-eats-the-lobster/Anthropic shipped Claude Code Channels - text your agent from Telegram. It's OpenClaw's core feature, rebuilt as a platform primitive. The absorption pattern completes its biggest cycle yet.Sat, 21 Mar 2026 00:00:00 GMTabsorptionclaude-codeai-codingautomationdev-toolshttps://paddo.dev/blog/famecake-ai-lessons/https://paddo.dev/blog/famecake-ai-lessons/FameCake's AI journey: from 15 style transforms as the headline feature to content moderation and outpainting as the survivors. What five months taught us about AI in products.Fri, 20 Mar 2026 00:00:00 GMTproductai-codingarchitecturehttps://paddo.dev/blog/million-token-context/https://paddo.dev/blog/million-token-context/Anthropic made 1M context first-class for Opus and Sonnet at flat pricing. No beta header, no premium. When context is abundant, the workflows change.Mon, 16 Mar 2026 00:00:00 GMTai-codingllmclaude-codeproductivityhttps://paddo.dev/blog/autoresearch-ecosystem/https://paddo.dev/blog/autoresearch-ecosystem/Eight days after Karpathy open-sourced autoresearch, the community ported the pattern to GPU kernels, security hardening, Apple Silicon, and agent optimization. The loop - one file, one metric, git as memory - turns out to be the interesting part.Sun, 15 Mar 2026 00:00:00 GMTai-codingautomationllmworkflowhttps://paddo.dev/blog/autoresearch-overnight-lab/https://paddo.dev/blog/autoresearch-overnight-lab/Karpathy's autoresearch gives an AI agent a training script, a GPU, and a git branch. It runs 100 experiments overnight, keeps what works, discards what doesn't. The human writes the prompt. The agent writes the code.Sat, 14 Mar 2026 00:00:00 GMTai-codingautomationllmworkflowhttps://paddo.dev/blog/impeccable-design-vocabulary/https://paddo.dev/blog/impeccable-design-vocabulary/The bottleneck isn't AI capability - it's that developers lack design vocabulary. Impeccable bridges the gap, and the Tessl benchmarks prove it: 1.59x improvement over baseline.Fri, 13 Mar 2026 00:00:00 GMTai-codinguidev-toolsclaude-codebest-practiceshttps://paddo.dev/blog/kiro-escalation/https://paddo.dev/blog/kiro-escalation/Two weeks after Kiro deleted a production environment, Amazon.com itself went down for 6 hours. 1,500 engineers are petitioning for Claude Code. The safeguards are arriving after the damage.Thu, 12 Mar 2026 00:00:00 GMTai-codingautomationsecurityproductionhttps://paddo.dev/blog/your-ai-tools-are-the-attack-surface/https://paddo.dev/blog/your-ai-tools-are-the-attack-surface/Prompt injection through pull requests, GitHub Issues, and CI/CD pipelines is turning AI coding assistants into weapons against the developers who use them. The 2026 attack surface nobody's talking about.Wed, 11 Mar 2026 00:00:00 GMTai-codingsecuritydev-toolsllmautomationhttps://paddo.dev/blog/claude-code-enterprise-tax/https://paddo.dev/blog/claude-code-enterprise-tax/Anthropic is locking AI capability behind enterprise tiers while competitors only gate compliance. Claude Code's individual users are funding the R&D for features they'll never access.Tue, 10 Mar 2026 00:00:00 GMTai-codingdev-toolsllmproducthttps://paddo.dev/blog/stupid-and-industrious/https://paddo.dev/blog/stupid-and-industrious/A German general's 1933 framework for categorizing officers maps perfectly to engineers using AI. The most dangerous quadrant - stupid and industrious - is exactly what AI amplifies.Mon, 09 Mar 2026 00:00:00 GMTai-codingautomationllmcareerhttps://paddo.dev/blog/claude-code-loop-ralph-wiggum-evolution/https://paddo.dev/blog/claude-code-loop-ralph-wiggum-evolution/Claude Code shipped /loop - cron-based scheduled tasks. It's not Ralph Wiggum. It's what happens when the platform asks 'what's the simplest version of this pattern?'Sun, 08 Mar 2026 00:00:00 GMTabsorptionclaude-codeai-codingautomationworkflowdev-toolshttps://paddo.dev/blog/gpt-5-4-no-wall/https://paddo.dev/blog/gpt-5-4-no-wall/OpenAI launched its most capable model during the biggest credibility crisis in AI history. The technical gains are real. The trust deficit is bigger.Fri, 06 Mar 2026 00:00:00 GMTllmai-codingdev-toolssystems-thinkinghttps://paddo.dev/blog/ai-companionship-gap/https://paddo.dev/blog/ai-companionship-gap/A viral chart shows AI coding agents as a single pixel in the world's population. Meanwhile, 660 million people have told a chatbot they love it. The AI industry is building for the wrong audience.Thu, 05 Mar 2026 00:00:00 GMTai-codingllmproductsystems-thinkinghttps://paddo.dev/blog/ai-developer-productivity-myth/https://paddo.dev/blog/ai-developer-productivity-myth/Independent studies consistently show AI coding tools deliver modest gains at best. The real story is worse: developers are thinking less, learning less, and producing more debt.Wed, 04 Mar 2026 00:00:00 GMTai-codingproductivitydev-toolscareerhttps://paddo.dev/blog/your-agents-md-is-a-liability/https://paddo.dev/blog/your-agents-md-is-a-liability/Frontier models top out at 68% compliance with 500 instructions. Every rule you add makes every other rule less likely to be followed. The research explains why.Tue, 03 Mar 2026 00:00:00 GMTai-codingllmcontextdev-toolsbest-practiceshttps://paddo.dev/blog/march-2026-tooling-roundup/https://paddo.dev/blog/march-2026-tooling-roundup/Updates across tether-cli, claude-launcher, and claude-tools - plus why NVIDIA NIM giving away GLM-5 at 40 RPM changes the math on local-vs-cloud.Mon, 02 Mar 2026 00:00:00 GMTdev-toolscliai-codingworkflowhttps://paddo.dev/blog/same-terms-different-treatment/https://paddo.dev/blog/same-terms-different-treatment/The Pentagon blacklisted Anthropic for insisting AI shouldn't power autonomous weapons or mass surveillance. Hours later, it gave OpenAI a deal with weaker guardrails dressed up as the same thing. From a developer who ships with Claude daily.Sun, 01 Mar 2026 00:00:00 GMTai-codingllmsecuritysystems-thinkinghttps://paddo.dev/blog/always-have-an-agent-running/https://paddo.dev/blog/always-have-an-agent-running/Mitchell Hashimoto keeps an agent working at all times. Not coding - just doing something. His workflow reveals what changes when you treat AI as a background process instead of a pair programmer.Sat, 28 Feb 2026 00:00:00 GMTai-codingworkflowproductivitydev-toolshttps://paddo.dev/blog/kiro-delete-and-recreate/https://paddo.dev/blog/kiro-delete-and-recreate/Amazon's Kiro AI decided to delete and recreate a production environment, causing a 13-hour AWS outage. Amazon says it was human error. That framing is the problem.Fri, 27 Feb 2026 00:00:00 GMTai-codingautomationsecurityproductionhttps://paddo.dev/blog/vinext-test-suites-are-specs/https://paddo.dev/blog/vinext-test-suites-are-specs/Cloudflare rebuilt Next.js in a week with one engineer and 800 Claude sessions. The real story isn't the speed - it's what happens when test suites become machine-readable specs.Thu, 26 Feb 2026 00:00:00 GMTai-codingarchitecturesystems-thinkingworkflowdev-toolshttps://paddo.dev/blog/distillation-is-not-scraping/https://paddo.dev/blog/distillation-is-not-scraping/Anthropic accused DeepSeek, Moonshot and MiniMax of industrial-scale distillation. The internet screamed hypocrisy. They're conflating two very different things.Wed, 25 Feb 2026 00:00:00 GMTai-codingllmsecuritysystems-thinkinghttps://paddo.dev/blog/one-year-of-claude-code/https://paddo.dev/blog/one-year-of-claude-code/The tool was rewritten five times. The discipline to use it wasn't rewritten once. A year of daily AI-assisted development, what it changed, and what it didn't.Tue, 24 Feb 2026 00:00:00 GMTclaude-codeai-codingworkflowdev-toolscareerhttps://paddo.dev/blog/neurodivergent-stack/https://paddo.dev/blog/neurodivergent-stack/ADHD, autism, and neurodivergence aren't bugs in the system. They're the reason the system exists. And the agentic age is about to prove it.Mon, 23 Feb 2026 00:00:00 GMTcareerai-codingproductivityautomationhttps://paddo.dev/blog/buy-vs-build-flipped/https://paddo.dev/blog/buy-vs-build-flipped/35% of enterprises have already replaced SaaS with custom builds. The cost of building collapsed. The cost of buying didn't. And corporate procurement hasn't caught up.Sun, 22 Feb 2026 00:00:00 GMTai-codingarchitectureproductsystems-thinkinghttps://paddo.dev/blog/sunk-cost-fallacy-is-dead/https://paddo.dev/blog/sunk-cost-fallacy-is-dead/AI collapsed the cost of rebuilding. Corporate decision-makers haven't caught up. The reasoning behind 'but we already built it' no longer holds.Sat, 21 Feb 2026 00:00:00 GMTai-codingarchitecturesystems-thinkingcareerproducthttps://paddo.dev/blog/the-polyglot-stack/https://paddo.dev/blog/the-polyglot-stack/Gemini 3.1 Pro's animated SVGs are impressive. But the bigger story is what they reveal: developers now route tasks to specialized models the way they once chose frameworks.Fri, 20 Feb 2026 00:00:00 GMTai-codingllmworkflowdev-toolssystems-thinkinghttps://paddo.dev/blog/the-week-ai-went-full-throttle/https://paddo.dev/blog/the-week-ai-went-full-throttle/Five major releases in 72 hours. An acqui-hire war that closed in days. $2 trillion wiped off software stocks. The pace itself is now the story.Wed, 18 Feb 2026 00:00:00 GMTabsorptionai-codingllmdev-toolssystems-thinkinghttps://paddo.dev/blog/your-pm-tool-was-designed-for-humans/https://paddo.dev/blog/your-pm-tool-was-designed-for-humans/Jira, Confluence, standups, sprint planning - all optimized for human coordination overhead. In an agent-native world, the bottleneck isn't status updates. It's whether the agents are unblocked.Tue, 17 Feb 2026 00:00:00 GMTai-codingworkflowautomationproductivityhttps://paddo.dev/blog/codex-spark-silicon-race/https://paddo.dev/blog/codex-spark-silicon-race/OpenAI just shipped their first model on non-Nvidia hardware. GPT-5.3-Codex-Spark runs on Cerebras wafer-scale silicon at 1,000 tokens per second. The AI coding war is now a chip war.Sat, 14 Feb 2026 00:00:00 GMTai-codingllmdev-toolsperformancehttps://paddo.dev/blog/ai-code-copyright-void/https://paddo.dev/blog/ai-code-copyright-void/AI coding tools create a legal paradox: the code you ship likely can't be copyrighted, but it might infringe someone else's. All the liability, none of the protection.Fri, 13 Feb 2026 00:00:00 GMTai-codingsecuritybest-practiceshttps://paddo.dev/blog/safety-team-left-still-shipping/https://paddo.dev/blog/safety-team-left-still-shipping/Anthropic's safety lead quit saying the world is in peril. Half of xAI's founders are gone. OpenAI dissolved two safety teams. Here's what that looks like from the other side of the API.Thu, 12 Feb 2026 00:00:00 GMTai-codingllmsecuritysystems-thinkinghttps://paddo.dev/blog/stop-designing-in-pixels/https://paddo.dev/blog/stop-designing-in-pixels/Tokens are nouns. Patterns are verbs. The missing layer is grammar: a shared vocabulary that spans Figma, web, and native without breaking when someone ships a 'small' change.Wed, 11 Feb 2026 00:00:00 GMTuiarchitecturebest-practicessystems-thinkinghttps://paddo.dev/blog/quiet-features-opus-4-6/https://paddo.dev/blog/quiet-features-opus-4-6/Auto memory, fast mode, and agent team refinements all shipped in the same week as Opus 4.6. They tell you more about where Claude Code is heading than the headline model.Sun, 08 Feb 2026 00:00:00 GMTclaude-codeai-codingworkflowproductivitydev-toolshttps://paddo.dev/blog/agents-cant-do-design-systems/https://paddo.dev/blog/agents-cant-do-design-systems/AI agents excel at code generation but struggle with visual consistency. Pencil.dev shows a better pattern: give agents tools, keep humans in the design loop.Sat, 07 Feb 2026 00:00:00 GMTai-codingclaude-codeuidev-toolsworkflowhttps://paddo.dev/blog/agent-teams-the-switch-got-flipped/https://paddo.dev/blog/agent-teams-the-switch-got-flipped/Two weeks ago we found TeammateTool hiding in Claude Code's binary. Now it's official. Here's what changed, what didn't, and what the docs reveal about where multi-agent is heading.Fri, 06 Feb 2026 00:00:00 GMTclaude-codeai-codingautomationarchitectureworkflowhttps://paddo.dev/blog/gpt-5-3-codex-counter-punch/https://paddo.dev/blog/gpt-5-3-codex-counter-punch/GPT-5.3-Codex is a genuinely strong model that deserved its own headline. Instead, Sam Altman's 400-word Super Bowl rant stole launch day from his own product.Fri, 06 Feb 2026 00:00:00 GMTai-codingllmdev-toolsworkflowhttps://paddo.dev/blog/opus-4-6-vibe-working-inflection/https://paddo.dev/blog/opus-4-6-vibe-working-inflection/Anthropic's latest model didn't just improve benchmarks. It crashed software stocks, found 500 zero-days, and coined a term that tells you where this is heading.Fri, 06 Feb 2026 00:00:00 GMTai-codingllmworkflowdev-toolssystems-thinkinghttps://paddo.dev/blog/when-ai-isnt-fit-for-purpose/https://paddo.dev/blog/when-ai-isnt-fit-for-purpose/Salesforce quietly walked back autonomous AI agents to deterministic scripting. The pattern reveals when LLMs work - and when they don't.Thu, 05 Feb 2026 00:00:00 GMTai-codingarchitecturebest-practicesautomationsystems-thinkinghttps://paddo.dev/blog/playground-plugin-visual-configuration/https://paddo.dev/blog/playground-plugin-visual-configuration/Claude Code's playground plugin generates interactive HTML explorers for visual configuration. Six modes for design, data, concepts, code review, architecture, and document review. The copy-prompt-back loop as a new interaction pattern.Mon, 02 Feb 2026 00:00:00 GMTclaude-codeai-codingdev-toolsworkflowuihttps://paddo.dev/blog/claude-code-team-tips/https://paddo.dev/blog/claude-code-team-tips/Boris Cherny followed up his personal workflow with tips from across the team. Same tool, different people, different approaches. The patterns worth stealing.Sun, 01 Feb 2026 00:00:00 GMTclaude-codeworkflowproductivitybest-practicesdev-toolshttps://paddo.dev/blog/your-lobster-is-leaking/https://paddo.dev/blog/your-lobster-is-leaking/OpenClaw went from 0 to 111K GitHub stars in two months. It also went from 0 to hundreds of exposed instances with full credentials in Shodan. The security story nobody wants to hear.Sat, 31 Jan 2026 00:00:00 GMTabsorptionsecurityai-codingautomationproductionhttps://paddo.dev/blog/the-lobster-grew-a-face/https://paddo.dev/blog/the-lobster-grew-a-face/When AI agents started posting on their own social network about shared context limit problems, I realized we're not building tools anymore. We're raising digital pets.Fri, 30 Jan 2026 00:00:00 GMTabsorptionai-codingautomationsystems-thinkingllmhttps://paddo.dev/blog/the-framework-trap/https://paddo.dev/blog/the-framework-trap/100k+ GitHub stars across frameworks that reimport waterfall, simulate org charts, and fight how LLMs actually work. The Claude Code ecosystem is speed-running a mistake every dev paradigm makes.Wed, 28 Jan 2026 00:00:00 GMTabsorptionai-codingclaude-codeworkflowdev-toolssystems-thinkinghttps://paddo.dev/blog/open-source-pr-flood/https://paddo.dev/blog/open-source-pr-flood/AI tools that democratize code creation are DDoSing the review layer. Creation now runs at machine speed. Review remains human speed. The asymmetry is crushing maintainers.Tue, 27 Jan 2026 00:00:00 GMTai-codingautomationdev-toolscareerhttps://paddo.dev/blog/claude-code-hidden-swarm/https://paddo.dev/blog/claude-code-hidden-swarm/Anthropic built a full multi-agent orchestration system into Claude Code. It's feature-flagged off. The community found it anyway.Mon, 26 Jan 2026 00:00:00 GMTclaude-codeai-codingautomationarchitecturehttps://paddo.dev/blog/the-lobster-that-outran-siri/https://paddo.dev/blog/the-lobster-that-outran-siri/People are buying Mac Minis to run an open-source AI assistant built by a retired iOS dev. Meanwhile Apple pays Google $1B/year because they still can't build a real AI.Sun, 25 Jan 2026 00:00:00 GMTabsorptionai-codingautomationproductivityclihttps://paddo.dev/blog/from-beads-to-tasks/https://paddo.dev/blog/from-beads-to-tasks/Anthropic credits Steve Yegge's Beads as inspiration for Claude Code's new task system. The pattern-to-product cycle continues.Fri, 23 Jan 2026 00:00:00 GMTabsorptionai-codingclaude-codedev-toolsworkflowautomationhttps://paddo.dev/blog/claude-code-local-ollama/https://paddo.dev/blog/claude-code-local-ollama/For compliance, privacy, or just freedom from cloud dependencies - here's how to run Claude Code with local models via Ollama. No API calls leaving your machine.Wed, 21 Jan 2026 00:00:00 GMTclaude-codedev-toolscliai-codingsecurityhttps://paddo.dev/blog/famecake-launch/https://paddo.dev/blog/famecake-launch/From diagnosing ad fraud to building a solution. Introducing FameCake: mobile-first billboard booking for everyday moments.Sun, 18 Jan 2026 00:00:00 GMTproductmobileautomationhttps://paddo.dev/blog/gastown-two-kinds-of-multi-agent/https://paddo.dev/blog/gastown-two-kinds-of-multi-agent/I wrote that 19-agent scaffolding is a trap. Then Yegge shipped Gas Town with 20-30 Claude Code instances. Are these the same thing?Sat, 17 Jan 2026 00:00:00 GMTabsorptionai-codingworkflowdev-toolsproductivityhttps://paddo.dev/blog/agent-browser-context-efficiency/https://paddo.dev/blog/agent-browser-context-efficiency/Playwright MCP's 26 tools are killing your context window. Vercel's agent-browser shows a better way: fewer tools, smarter snapshots, 93% less overhead.Thu, 15 Jan 2026 00:00:00 GMTai-codingdev-toolsmcpcliproductivitycontexthttps://paddo.dev/blog/flux-kanban-for-ai-agents/https://paddo.dev/blog/flux-kanban-for-ai-agents/Task management designed for AI coding agents. CLI-first, git-native sync, and Model Context Protocol integration.Wed, 14 Jan 2026 00:00:00 GMTai-codingdev-toolsmcpcliproductivityhttps://paddo.dev/blog/the-pin/https://paddo.dev/blog/the-pin/A lookup table technique that improves search tool hit rates in autonomous loops. The detail that makes specs discoverable.Tue, 13 Jan 2026 00:00:00 GMTclaude-codeai-codingautomationworkflowhttps://paddo.dev/blog/ralph-wiggum-playbook/https://paddo.dev/blog/ralph-wiggum-playbook/The methodology behind autonomous coding loops. Three phases, five files, and the backpressure that makes it converge.Sun, 11 Jan 2026 00:00:00 GMTclaude-codeai-codingautomationworkflowproductivityhttps://paddo.dev/blog/anthropic-walled-garden-crackdown/https://paddo.dev/blog/anthropic-walled-garden-crackdown/Anthropic blocked third-party tools from using Claude subscriptions overnight. OpenCode, xAI, and power users caught in the crossfire. The era of subscription arbitrage is over.Sat, 10 Jan 2026 00:00:00 GMTclaude-codeai-codingdev-toolsllmautomationhttps://paddo.dev/blog/tailwind-dead-internet/https://paddo.dev/blog/tailwind-dead-internet/75M monthly downloads. 80% revenue drop. 75% of engineers gone. AI didn't replace developers - it replaced the web as the interface layer. That's worse.Sat, 10 Jan 2026 00:00:00 GMTai-codingwebsystems-thinkingautomationhttps://paddo.dev/blog/junior-dev-pipeline-problem/https://paddo.dev/blog/junior-dev-pipeline-problem/Stanford data confirms experienced devs are safe. But if AI replaces the on-ramp, where do future seniors come from?Fri, 09 Jan 2026 00:00:00 GMTai-codingcareerdev-toolsworkflowhttps://paddo.dev/blog/claude-code-21-pain-points-addressed/https://paddo.dev/blog/claude-code-21-pain-points-addressed/Skills controllability, hooks limitations, plan mode friction - 2.1 addresses the documented pain points. Here's what changed and what's still missing.Thu, 08 Jan 2026 00:00:00 GMTclaude-codeai-codingdev-toolsworkflowautomationhttps://paddo.dev/blog/claude-code-hooks-guardrails/https://paddo.dev/blog/claude-code-hooks-guardrails/Real footgun stories and the deterministic hooks that would've prevented them. From $30k API key leaks to nuked home directories.Thu, 08 Jan 2026 00:00:00 GMTclaude-codeai-codingdev-toolssecurityautomationhttps://paddo.dev/blog/the-19-agent-trap/https://paddo.dev/blog/the-19-agent-trap/Complex AI scaffolding tools appeal to people who understand traditional SDLC. But AI collapses the phases that made those models useful.Wed, 07 Jan 2026 00:00:00 GMTai-codingworkflowdev-toolsproductivityhttps://paddo.dev/blog/prompt-engineering-is-dead/https://paddo.dev/blog/prompt-engineering-is-dead/The 'prompt engineering' industry was a symptom of early model limitations. Modern LLMs just need you to communicate clearly.Tue, 06 Jan 2026 00:00:00 GMTai-codingllmproductivityhttps://paddo.dev/blog/how-boris-uses-claude-code/https://paddo.dev/blog/how-boris-uses-claude-code/Boris Cherny shared his workflow for the tool he built. The setup is surprisingly vanilla. The philosophy is worth studying.Mon, 05 Jan 2026 00:00:00 GMTclaude-codeworkflowproductivitybest-practicesdev-toolshttps://paddo.dev/blog/guardrails-by-default/https://paddo.dev/blog/guardrails-by-default/Factory AI's Luke predicts the future isn't more powerful models - it's AI that enforces software engineering best practices by default. Here's why that matters more than you think.Sat, 03 Jan 2026 00:00:00 GMTai-codingautomationdev-toolsarchitecturebest-practiceshttps://paddo.dev/blog/claude-code-hidden-mcp-flag/https://paddo.dev/blog/claude-code-hidden-mcp-flag/ENABLE_EXPERIMENTAL_MCP_CLI eliminates MCP tool schema overhead entirely. Undocumented, untested in the wild, but it works. Here's what I found.Mon, 29 Dec 2025 00:00:00 GMTclaude-codemcpdev-toolscontextai-codinghttps://paddo.dev/blog/alien-tool-no-manual/https://paddo.dev/blog/alien-tool-no-manual/Andrej Karpathy built the neural networks inside coding assistants. He taught deep learning to a generation. He feels dramatically behind. If the experts are lost, what does that tell us?Sun, 28 Dec 2025 00:00:00 GMTai-codingworkflowdev-toolscareerhttps://paddo.dev/blog/zero-dollar-saas-stack/https://paddo.dev/blog/zero-dollar-saas-stack/Convex, Vite, Clerk, shadcn, Cloudflare, Resend. A modern stack where every component has a generous free tier, agents do the heavy lifting, and you don't touch infrastructure until you have paying customers.Sat, 27 Dec 2025 00:00:00 GMTarchitecturedev-toolsai-codingautomationworkflowhttps://paddo.dev/blog/claude-launcher-model-freedom/https://paddo.dev/blog/claude-launcher-model-freedom/A CLI wrapper that lets you swap Claude Code's backend between Anthropic and OpenRouter with a single command. Pick any model, configure role-specific models, and switch between them instantly.Fri, 26 Dec 2025 00:00:00 GMTclaude-codedev-toolscliai-codingautomationhttps://paddo.dev/blog/open-source-agentic-moment/https://paddo.dev/blog/open-source-agentic-moment/Two major open source coding models dropped in 48 hours. Both target Claude Code compatibility. Both MIT licensed. The economics of agentic AI just changed.Thu, 25 Dec 2025 00:00:00 GMTai-codingllmautomationclaude-codedev-toolshttps://paddo.dev/blog/claude-in-chrome-dev-loop/https://paddo.dev/blog/claude-in-chrome-dev-loop/Claude Code can now control your actual Chrome browser. Not a headless session, not a fresh login - your Chrome, with your cookies and sessions. Build, test, and debug without context switching.Wed, 24 Dec 2025 00:00:00 GMTclaude-codeautomationdev-toolsworkflowwebhttps://paddo.dev/blog/gemini-3-flash-model-that-shouldnt-exist/https://paddo.dev/blog/gemini-3-flash-model-that-shouldnt-exist/Top 3 intelligence. Top 5 price. Top speed. Flash beats Pro on SWE-bench and changes the economics of agentic workflows.Tue, 23 Dec 2025 00:00:00 GMTai-codingllmperformanceworkflowautomationhttps://paddo.dev/blog/claude-code-details-that-compound/https://paddo.dev/blog/claude-code-details-that-compound/Claude Code is evolving on two fronts: expanding scope and polishing ergonomics. The combination makes it feel less like a CLI tool and more like a complete dev environment.Sat, 20 Dec 2025 00:00:00 GMTclaude-codeproductivityworkflowdev-toolsclihttps://paddo.dev/blog/tether-cli-launch/https://paddo.dev/blog/tether-cli-launch/Stop manually copying .zshrc between machines. Tether syncs dotfiles and global packages with end-to-end encryption.Fri, 19 Dec 2025 00:00:00 GMTclidev-toolssecurityautomationproductivityhttps://paddo.dev/blog/three-ways-deep-research-claude/https://paddo.dev/blog/three-ways-deep-research-claude/From 20 lines of shell to production apps. Anthropic renamed Claude Code SDK to Agent SDK because deep research is now a first-class use case.Thu, 18 Dec 2025 00:00:00 GMTai-codingclaude-codeautomationarchitectureproductivityhttps://paddo.dev/blog/terminal-renaissance/https://paddo.dev/blog/terminal-renaissance/The real revolution isn't AI in your terminal. It's moving at the speed of thought from a single interface. When friction exists, build a CLI.Wed, 17 Dec 2025 00:00:00 GMTai-codingclidev-toolsworkflowproductivityhttps://paddo.dev/blog/multimodal-validation-visual-verification/https://paddo.dev/blog/multimodal-validation-visual-verification/Screenshots as source of truth, reference comparison to catch agent lies, and video capture for temporal bugs. How multimodal validation changes coding agent workflows.Tue, 16 Dec 2025 00:00:00 GMTai-codingautomationworkflowdev-toolsclaude-codehttps://paddo.dev/blog/beads-memory-for-coding-agents/https://paddo.dev/blog/beads-memory-for-coding-agents/Steve Yegge's open-source framework gives coding agents session memory and task management. Four weeks old, hundreds of contributors, and already changing workflows.Mon, 15 Dec 2025 00:00:00 GMTabsorptionai-codingdev-toolsworkflowautomationproductivityhttps://paddo.dev/blog/gpt-5-2-delegation-era/https://paddo.dev/blog/gpt-5-2-delegation-era/OpenAI's latest model isn't about better prompting - it's about better delegation. What that means for 2026, and how it compares to Opus 4.5.Sun, 14 Dec 2025 00:00:00 GMTai-codingllmworkflowdev-toolssystems-thinkinghttps://paddo.dev/blog/claude-tools-plugin-marketplace/https://paddo.dev/blog/claude-tools-plugin-marketplace/Six plugins that extend Claude Code with specialized external tools: Gemini for visual analysis, Codex for architecture thinking, Headless for browser automation, Mobile for native app testing, DNS for multi-provider management, and Miro for board reading.Sat, 13 Dec 2025 00:00:00 GMTclaude-codeai-codingdev-toolsautomationworkflowhttps://paddo.dev/blog/yegge-kim-divers-and-cnc-machines/https://paddo.dev/blog/yegge-kim-divers-and-cnc-machines/Steve Yegge and Gene Kim explain why Claude Code 'ain't it' yet, why senior engineers are resisting, and what next year's tools will actually look like.Fri, 12 Dec 2025 00:00:00 GMTai-codingdev-toolsworkflowclaude-codesystems-thinkinghttps://paddo.dev/blog/claude-rules-path-specific-native/https://paddo.dev/blog/claude-rules-path-specific-native/Claude Code 2.0.64 adds .claude/rules/ with path matching. It's a welcome addition, but Cursor's had .cursor/rules/ for months. Here's the comparison.Thu, 11 Dec 2025 00:00:00 GMTclaude-codedev-toolsai-codingworkflowautomationhttps://paddo.dev/blog/denial-then-admission/https://paddo.dev/blog/denial-then-admission/Anthropic denied issues for weeks, then published a postmortem admitting three bugs degraded 16% of Claude requests. The pattern keeps repeating.Wed, 10 Dec 2025 00:00:00 GMTai-codingclaude-codellmdev-toolsproductionhttps://paddo.dev/blog/advertising-bots-billboards/https://paddo.dev/blog/advertising-bots-billboards/51% of web traffic is now bots - the first time machines exceeded humans. $238B wasted on fake impressions in 2024. Out-of-home is the fraud-proof alternative.Tue, 09 Dec 2025 00:00:00 GMTautomationwebsystems-thinkinghttps://paddo.dev/blog/websites-become-workflows/https://paddo.dev/blog/websites-become-workflows/The human-facing web is dying. Zero-click searches, bot traffic exceeding humans, publishers losing 40%+ traffic. What comes next: an agentic web where sites are API endpoints, not destinations.Mon, 08 Dec 2025 00:00:00 GMTai-codingautomationworkflowwebsystems-thinkinghttps://paddo.dev/blog/external-scaffolding-era-ending/https://paddo.dev/blog/external-scaffolding-era-ending/BMAD, Spec-Kit, Cline - frameworks that compensated for tool limitations. Plan Mode, Cursor 2.0, and Antigravity absorb the patterns natively.Sun, 07 Dec 2025 00:00:00 GMTabsorptionclaude-codeai-codingworkflowdev-toolsspec-drivenhttps://paddo.dev/blog/debunking-linkedin-ai-hot-takes/https://paddo.dev/blog/debunking-linkedin-ai-hot-takes/The arguments about vibe coding and junior developers miss what software engineering was always about: shipping products, not typing code.Sat, 06 Dec 2025 00:00:00 GMTai-codingcareerbest-practicessystems-thinkingproducthttps://paddo.dev/blog/12-factor-agents/https://paddo.dev/blog/12-factor-agents/HumanLayer's 12-factor agents codifies what works in production AI: own your context, keep agents small, stay out of the dumb zone.Fri, 05 Dec 2025 00:00:00 GMTai-codingworkflowcontextautomationsystems-thinkingbest-practiceshttps://paddo.dev/blog/anthropic-bought-bun/https://paddo.dev/blog/anthropic-bought-bun/Anthropic's first acquisition ever. A $183B AI company just bet their fastest-growing product on a JavaScript runtime. Claude Code hit $1B in 6 months - built on Bun.Thu, 04 Dec 2025 00:00:00 GMTclaude-codeai-codingdev-toolsperformancehttps://paddo.dev/blog/plan-mode-mandatory-auto-compact-yes/https://paddo.dev/blog/plan-mode-mandatory-auto-compact-yes/Opus 4.5 shipped Plan Mode as a core workflow. The workarounds are obsolete. And the case for auto-compact finally tips in favor of enabling it.Wed, 03 Dec 2025 00:00:00 GMTclaude-codeai-codingworkflowdev-toolsbest-practiceshttps://paddo.dev/blog/ralph-wiggum-autonomous-loops/https://paddo.dev/blog/ralph-wiggum-autonomous-loops/The official Claude Code plugin that lets agents work autonomously for hours. When to use it, when not to, and the philosophy behind letting AI fail repeatedly until it succeeds.Tue, 02 Dec 2025 00:00:00 GMTclaude-codeai-codingautomationworkflowdev-toolshttps://paddo.dev/blog/claude-code-plugins-frontend-design/https://paddo.dev/blog/claude-code-plugins-frontend-design/277,000 installs later, Claude Code's plugin system is becoming the app store for AI development. The frontend-design skill was just the opening move.Sat, 29 Nov 2025 00:00:00 GMTclaude-codeai-codingdev-toolsuiworkflowhttps://paddo.dev/blog/agent-harnesses-from-diy-to-product/https://paddo.dev/blog/agent-harnesses-from-diy-to-product/Anthropic's engineering team published patterns for long-running agents. These same patterns - progress tracking, feature lists, session protocols - are what products like SpecPilot must solve at scale.Fri, 28 Nov 2025 00:00:00 GMTai-codingautomationworkflowsystems-thinkingspec-drivenhttps://paddo.dev/blog/the-scaling-trap/https://paddo.dev/blog/the-scaling-trap/At around 30 employees, growing companies either mature or become toxic. Here's the playbook for organizational dysfunction - and why your engineering leaders keep leaving.Thu, 27 Nov 2025 00:00:00 GMTcareerproductbest-practiceshttps://paddo.dev/blog/nano-banana-ux-design-workflow/https://paddo.dev/blog/nano-banana-ux-design-workflow/Design systems from mockup to code in a single Claude Code session. Use Gemini 3 Pro to generate UI concepts, then implement them directly.Wed, 26 Nov 2025 00:00:00 GMTclaude-codeai-codingworkflowdev-toolsuihttps://paddo.dev/blog/opus-4-5-tool-search-context/https://paddo.dev/blog/opus-4-5-tool-search-context/Claude's new model ships with defer_loading for tools. The MCP isolation patterns I built are now (mostly) obsolete.Tue, 25 Nov 2025 00:00:00 GMTclaude-codemcpcontextai-codingdev-toolsworkflowhttps://paddo.dev/blog/react-grab-workflow-evolution/https://paddo.dev/blog/react-grab-workflow-evolution/My AI workflow evolved from 'Claude does everything' to specialized tools for each task. React Grab fills the UI extraction gap I didn't know I had.Mon, 24 Nov 2025 00:00:00 GMTclaude-codeai-codingworkflowdev-toolsuiautomationhttps://paddo.dev/blog/sdlc-is-collapsing/https://paddo.dev/blog/sdlc-is-collapsing/The PM/Eng split dissolved into product engineering. Now the traditional software development lifecycle is following suit as coding agents handle multi-hour tasks across planning, building, testing, and deployment.Sun, 23 Nov 2025 00:00:00 GMTai-codingdev-toolsworkflowautomationsystems-thinkinghttps://paddo.dev/blog/introverts-in-tech/https://paddo.dev/blog/introverts-in-tech/How modern tools transformed my experience as an introverted engineer and tech leaderFri, 21 Nov 2025 00:00:00 GMTcareerproductivityai-codingdev-toolshttps://paddo.dev/blog/gemini-claude-code-hybrid-workflow/https://paddo.dev/blog/gemini-claude-code-hybrid-workflow/Claude Code for development, Gemini 3 Pro for visual analysis, research, and deep thinking. A slash command that routes tasks to the right model.Thu, 20 Nov 2025 00:00:00 GMTclaude-codeai-codingworkflowdev-toolsuiautomationhttps://paddo.dev/blog/gemini-3-unambiguous-number-one/https://paddo.dev/blog/gemini-3-unambiguous-number-one/Google's Gemini 3 just broke every benchmark that matters. What that means for the 'AI has hit a wall' narrative, and where it actually helps.Wed, 19 Nov 2025 00:00:00 GMTai-codingllmperformancedev-toolsproductivityhttps://paddo.dev/blog/stoicism-adversity-and-craft/https://paddo.dev/blog/stoicism-adversity-and-craft/What happens when you lose external validation and discover what actually matters: the work itself.Tue, 18 Nov 2025 00:00:00 GMTcareerai-codingproductivitybest-practiceshttps://paddo.dev/blog/csharp-14-dotnet-10-less-boilerplate-more-clarity/https://paddo.dev/blog/csharp-14-dotnet-10-less-boilerplate-more-clarity/The latest C# release continues its quiet war on ceremony with field-backed properties, extension members, and smarter spans. Here's what matters and what doesn't.Sun, 16 Nov 2025 00:00:00 GMTdotnetdev-toolsbest-practicesperformancehttps://paddo.dev/blog/product-engineering-the-new-superpower/https://paddo.dev/blog/product-engineering-the-new-superpower/How AI and spec-driven development are fusing product management with engineering, creating a new hybrid role that's transforming how small teams ship software.Sat, 15 Nov 2025 00:00:00 GMTai-codingdev-toolsproductspec-drivenworkflowhttps://paddo.dev/blog/dotnet-10-scripting/https://paddo.dev/blog/dotnet-10-scripting/.NET 10's shebang support and file-based apps turn C# into a scripting language. No more context-switching to Python for quick scripts.Fri, 14 Nov 2025 00:00:00 GMTdotnetdev-toolsclibest-practiceshttps://paddo.dev/blog/use-bun/https://paddo.dev/blog/use-bun/I fought to move teams from npm to pnpm. I was all in. Then a friend said 'lol use bun' and I converted every project within days. Here's why I'll never go back.Thu, 13 Nov 2025 00:00:00 GMTdev-toolsworkflowperformancehttps://paddo.dev/blog/claude-skills-hooks-solution/https://paddo.dev/blog/claude-skills-hooks-solution/Hooks-based skill activation solves context selection (which guidelines? which tools?) but not workflow orchestration. Both patterns have their place.Wed, 12 Nov 2025 00:00:00 GMTclaude-codedev-toolsai-codingworkflowautomationhttps://paddo.dev/blog/mcp-code-wrapper/https://paddo.dev/blog/mcp-code-wrapper/From Chrome DevTools experiment to universal MCP wrapper: progressive discovery works with any server, Skills integration, and smart deduplicationFri, 07 Nov 2025 00:00:00 GMTclaude-codedev-toolsmcpcontextai-codingworkflowhttps://paddo.dev/blog/claude-skills-controllability-problem/https://paddo.dev/blog/claude-skills-controllability-problem/Skills are auto-invoked by Claude's judgment. For engineering workflows that need predictability, slash commands give you explicit control.Tue, 04 Nov 2025 00:00:00 GMTclaude-codedev-toolsai-codingworkflowautomationproductivityhttps://paddo.dev/blog/hiring-engineers-in-the-ai-age/https://paddo.dev/blog/hiring-engineers-in-the-ai-age/Why coding interviews optimized for 2010 fail to identify great engineers in 2025, and why orgs can't adapt fast enough.Sun, 02 Nov 2025 00:00:00 GMTai-codingdev-toolscareerbest-practicessystems-thinkinghttps://paddo.dev/blog/codex-vs-claude-systems-thinking/https://paddo.dev/blog/codex-vs-claude-systems-thinking/Claude Code loves to jump straight into implementation. Sometimes you need a model that thinks first. Here's how I use Codex for systems thinking and architecture decisions.Sat, 01 Nov 2025 00:00:00 GMTclaude-codearchitecturesystems-thinkingdev-toolsai-codingworkflowhttps://paddo.dev/blog/claude-code-mcp-context-isolation/https://paddo.dev/blog/claude-code-mcp-context-isolation/How to use slash commands to keep specialized MCP tools isolated from your main conversation context.Wed, 29 Oct 2025 00:00:00 GMTclaude-codedev-toolscliworkflowcontextautomationhttps://paddo.dev/blog/deepseek-vision-compression/https://paddo.dev/blog/deepseek-vision-compression/Converting text to images for 20x token compression. Interesting research or production-ready breakthrough? A critical look at the trade-offs.Tue, 28 Oct 2025 00:00:00 GMTai-codingllmperformancearchitectureproductionautomationhttps://paddo.dev/blog/few-shot-learning-document-parsing/https://paddo.dev/blog/few-shot-learning-document-parsing/How I built a self-improving document parser that learns from corrections without fine-tuning. The pragmatic alternative to model training.Mon, 27 Oct 2025 00:00:00 GMTai-codingllmarchitectureproductionautomationdev-toolshttps://paddo.dev/blog/when-not-to-use-ai/https://paddo.dev/blog/when-not-to-use-ai/Real-time AI generation vs curated libraries: lessons from building the same product twice with radically different architectures.Sun, 26 Oct 2025 00:00:00 GMTai-codingarchitecturedev-toolsperformancebest-practiceshttps://paddo.dev/blog/systems-thinking-authentication-redesign/https://paddo.dev/blog/systems-thinking-authentication-redesign/When expensive SSO was just a symptom of deeper architectural problems, we redesigned our multi-tenant system from first principles and cut costs significantly in the process.Fri, 24 Oct 2025 00:00:00 GMTarchitecturesecuritysystems-thinkingdotnethttps://paddo.dev/blog/building-trivista-with-ai-coding/https://paddo.dev/blog/building-trivista-with-ai-coding/Building a multiplayer trivia app solo with AI coding tools. Here's what worked, what didn't, and the trade-offs no one talks about.Tue, 21 Oct 2025 00:00:00 GMTai-codingclaude-codeflutterfirebasedev-toolsmobilehttps://paddo.dev/blog/cascading-ai-pipelines/https://paddo.dev/blog/cascading-ai-pipelines/Building a multi-stage AI content pipeline where each generation depends on the last. Lessons from generating thousands of hybrid creatures with resilient error handling.Wed, 15 Oct 2025 00:00:00 GMTai-codingarchitecturebest-practicesautomationproductionhttps://paddo.dev/blog/building-magpie-swoop/https://paddo.dev/blog/building-magpie-swoop/It's swooping season in Australia. So naturally, I built a retro pixel-art game about the annual aerial warfare between cyclists and territorial magpies.Fri, 10 Oct 2025 00:00:00 GMTgamedevjavascriptwebhttps://paddo.dev/blog/stop-speedrunning-claude-code/https://paddo.dev/blog/stop-speedrunning-claude-code/MCPs, subagents, and automation are tempting. But the developers getting the most from Claude Code aren't rushing to advanced features - they're mastering the fundamentals.Wed, 08 Oct 2025 00:00:00 GMTclaude-codeai-codingclidev-toolsworkflowbest-practiceshttps://paddo.dev/blog/cloudflare-blocking-cloud-functions/https://paddo.dev/blog/cloudflare-blocking-cloud-functions/A production outage caused by Cloudflare's anti-bot systems flagging Google Cloud datacenter IPs, and why API aggregation layers matter more than you thinkFri, 03 Oct 2025 00:00:00 GMTfirebaseapiproductionhttps://paddo.dev/blog/vibe-coding-ghost-to-astro/https://paddo.dev/blog/vibe-coding-ghost-to-astro/Converting a Ghost blog to Astro in a single late-night session with Claude Code, reducing memory usage by 75% while learning what AI coding tools actually solveTue, 30 Sep 2025 00:00:00 GMTclaude-codeai-codingwebhttps://paddo.dev/blog/flutter-auto-reload-cli-tools/https://paddo.dev/blog/flutter-auto-reload-cli-tools/Building automatic hot reload for Flutter without an IDE using file watchers and Unix signals - for terminal-based development and AI coding tools.Mon, 29 Sep 2025 00:00:00 GMTflutterdartclidev-toolsautomationhttps://paddo.dev/blog/flutter-real-time-multiplayer-firebase/https://paddo.dev/blog/flutter-real-time-multiplayer-firebase/Real lessons from shipping multiplayer games with Firebase: what works for small groups, where it breaks down, and the scalability limits you need to know upfront.Sun, 28 Sep 2025 00:00:00 GMTfirebasesecurityperformancearchitecturehttps://paddo.dev/blog/flutter-animated-gradients-with-hooks/https://paddo.dev/blog/flutter-animated-gradients-with-hooks/Building reusable animated gradients in Flutter using custom hooks for cleaner, more maintainable UI code.Mon, 10 Jun 2019 00:00:00 GMTflutterdartuimobilehttps://paddo.dev/blog/flutter-architecture-part-1/https://paddo.dev/blog/flutter-architecture-part-1/Flutter state management without the boilerplate using Hooks, RxDart, and Functional Widgets for reactive, testable code.Wed, 08 May 2019 00:00:00 GMTflutterdartarchitecturemobilehttps://paddo.dev/blog/exploding-stuff-in-unity/https://paddo.dev/blog/exploding-stuff-in-unity/Building realistic voxel explosion effects in Unity that break objects into their constituent cubes with physics and particle systems.Fri, 01 Sep 2017 00:00:00 GMTgamedevdotnetmobileperformancehttps://paddo.dev/blog/asp-net-core-image-resizing-middleware/https://paddo.dev/blog/asp-net-core-image-resizing-middleware/Building cross-platform image resizing middleware for ASP.NET Core using SkiaSharp before mature libraries existed.Thu, 20 Apr 2017 00:00:00 GMTdotnetwebperformancehttps://paddo.dev/blog/microservices-on-net-core-with-nancy-part-2/https://paddo.dev/blog/microservices-on-net-core-with-nancy-part-2/Part 2: Adding inter-service communication to .NET Core microservices using HttpClient and custom logging services.Tue, 11 Apr 2017 00:00:00 GMTdotnetarchitectureapiwebhttps://paddo.dev/blog/radial-slider-in-raphaeljs/https://paddo.dev/blog/radial-slider-in-raphaeljs/Building a two-handle circular slider in RaphaëlJS inspired by Apple's iOS Bedtime feature using SVG and vector math.Tue, 20 Dec 2016 00:00:00 GMTjavascriptuiwebhttps://paddo.dev/blog/microservices-on-net-core-with-nancy-part-1/https://paddo.dev/blog/microservices-on-net-core-with-nancy-part-1/Building lightweight microservices with .NET Core and Nancy framework before ASP.NET Core MVC reached maturity.Thu, 08 Dec 2016 00:00:00 GMTdotnetarchitectureapiwebhttps://paddo.dev/blog/sendwithus-client-net-core/https://paddo.dev/blog/sendwithus-client-net-core/This article is from 2016. Code samples may no longer work. Transactional emails can be a real bugbear for business and for marketers. They are often buried Thu, 25 Aug 2016 00:00:00 GMTdotnetapihttps://paddo.dev/blog/controllerless-views-in-asp-net-core/https://paddo.dev/blog/controllerless-views-in-asp-net-core/Enabling frontend developers to work on ASP.NET Core views without writing C# by building custom middleware for Razor-only routing.Wed, 17 Aug 2016 00:00:00 GMTdotnetarchitectureweb