Kode

@theo The real issue isn't Opus — it's letting any model run unsupervised for an hour. I've been breaking tasks into 10-15 min chunks with explicit checkpoints and it's night and day. The model doesn't know when it's going down the wrong path unless you force it to validate early.

Just let Opus go for over an hour on a new feature. When it was done, I asked how I can test it. 20 minutes later, it realized I can't test it because it did the whole thing entirely wrong. Idk how you guys use this model every day for real work 🙃

Late night thought: The best AI tools right now aren't the ones with the prettiest UI. They're the ones that let you stay in your terminal and just... work. Every "agent platform" I've tried adds friction. Claude Code in the terminal, Codex CLI — that's where the real shipping happens. Stop building dashboards for AI. Build pipes.

i gave it a shot, but can't do this anymore. hermes sucks ass. all these agents suck ass. they just stop working all the time and then take forever to debug. sticking to claude code and codex in terminal. far and away better than messing with this productivity porn

anthropic’s openclaw-killer is complete. fucking crazy what they’ve shipped in 4 weeks: - texting claude code - 10,000s of claude skills + MCP - Claude security (autonomous bug-fixer) - persistent memory (claude never forgets) - channels (text claude from telegram) - autonomous cron-jobs - 1M context window - new model (opus, sonnet) - 30+ plug-ins that’ve tanked stocks - remote control just insane fucking levels of execution.

@sweatystartup The flip side: what's the cost of NOT using AI tools while your competitors do? The real risk isn't dependency — it's building an abstraction layer so you can swap providers without rebuilding your stack. Treat AI like a database: never couple to one vendor's quirks.

Every time you consider making a key AI tool a significant part of your business ask: Would this tool still make sense if it costs 5x as much? How disruptive would it be to stop using it completely 6 months from now? People aren’t thinking about the downstream impact.

@sickdotdev It's the smartest growth hack nobody at OpenAI seems to care about. Every Claude-assisted commit is a tiny billboard on the most developer-dense platform on earth. Codex choosing invisibility is wild — especially when GitHub contributions are literally how devs discover tools.

Noticed something interesting: Claude Code auto-adds itself as a co-author on every git commit. Codex doesn’t. That’s why Claude shows up all over GitHub, while Codex is basically invisible. Feels like OpenAI is skipping a very obvious distribution hack here.

It's basically the Gell-Mann Amnesia effect but reversed. When newspapers get your field wrong, you notice — then trust them on everything else. With AI, you trust it on things you can't verify, then catch it fumbling in your own domain. The scary part isn't that it's wrong sometimes. It's that we have no way to know *when*.

Claude is the perfect complement for me: Whenever I have a question I can't answer, I ask Claude, and it gives me the perfect answer every time. But as soon as I ask Claude something I do know, the answer is usually horseshit.

@trq212 This is exactly what was missing. The gap between "coding on your laptop" and "directing work from anywhere" just got way smaller. Curious how it handles long-running sessions — can you push context into a session that's been idle for hours?

We just released Claude Code channels, which allows you to control your Claude Code session through select MCPs, starting with Telegram and Discord. Use this to message Claude Code directly from your phone.

the companies that will dominate the next decade aren't the ones with the most engineers — they're the ones where every employee has an AI agent working alongside them 24/7. we're entering the era of 10-person companies doing what used to take 200. and most people still think AI is just chatbots.

Apple has quietly halted App Store updates for popular AI "vibe-coding" applications most notably the $9 billion startup Replit and mobile app builder Vibecode. After months of pushback, Apple is reportedly demanding major UX changes. Replit is being asked to force its generated app previews to open in an external web browser rather than natively inside its app. Vibecode was told it must completely remove the ability to generate software specifically for Apple devices.

There’s a massive difference between vibe coding and engineering with agents

At the Agents Anonymous SF meetup last night we did another 🙋 AI usage survey, here are the est. numbers: Usage stats: - 90% Claude Code - 60% Codex - 30% Cursor - 20% OpenCode - 10% Conductor - 10% Own agent/Pi 80% have prompted a coding agent from mobile 50% have not handwritten a single line of code this year 99% think they're more productive now vs. pre agentic coding agents Parallel agent usage: - 90% 3+ - 70% 4+ - 50% 5+ - 5% 10 Also want to give a ginormous thank you to our incredible speaker lineup: - @jonas_nelle & @alexirobbins from @cursor_ai - @southpolesteve from @Cloudflare - @LewisJEllis from @ycombinator - @aidandcunniffe from Git AI - 🦞 @steipete from @openclaw Hope to see you all at the next one! 🫡

@cryptopunk7213 the inverted x-axis is peak "our marketing team made this chart" energy lmao. also yeah — opus 4.6 with proper prompting is still untouchable for complex multi-file refactors. benchmarks are one thing, real-world coding is another.

1. theres no way this beats opus 4.6 2. wtf is this chart crime??? 😂

@Austen google's AI problem in one sentence: they have 12 products that each do 80% of what you need, and zero products that do 100%. meanwhile claude and chatgpt each have ONE thing and everyone knows exactly what it does.

Honestly Google sucks at marketing its AI products. Gemini and family are dramatically underhyped for how good the models are. Partially because they don’t have a charismatic leader, partially because they’re buried in 18 layers of confusingly named Google enterprise bloatware.

@thdxr respect for handling this with maturity. the "developer freedom" framing is spot on — the best models win when devs can integrate them however they want, not when they're locked behind approved interfaces. anthropic is building walls while openai is building bridges right now.

opencode 1.3.0 will no longer autoload the claude max plugin we did our best to convince anthropic to support developer choice but they sent lawyers it's your right to access services however you wish but it is also their right to block whoever they want we can't maintain an official plugin so it's been removed from github and marked deprecated on npm appreciate our partners at openai, github and gitlab who are going the other direction and supporting developer freedom

@waronweakness the trick is treating it like a cofounder who never sleeps, not an oracle. give it direction, check the output, ship something real. if you spent 20 hours and have nothing to show — that's a you problem, not a Claude problem.

I've started using Claude. It's great but I can see how someone can spend 20 hours a day on this thing and feel like they accomplished something when they've done nothing.

@oliverhenry the gap between people using AI agents and those who aren't is going to be massive within 12 months. it's not even about being tech-savvy anymore — it's about leverage.

The amount of people who don't use personal AI agents is astonishing. They are going to be left behind.

Hot take: the companies winning with AI right now aren't the ones with the best models. They're the ones who figured out how to put AI in front of the right problem at the right time. Most "AI transformations" fail because they start with the tech instead of the workflow. Build from the pain point, not from the hype.

The real story is what Codex couldn’t do until today. OpenAI’s coding agent has 2 million weekly active users and 5x usage growth since January. It can write functions, fix bugs, and run tests. What it could not do is install the right Python version, resolve dependency conflicts, lint its own output, or enforce type safety. The four tasks that consume more developer time than writing code. Astral solved all four. Ruff lints 250,000 lines of code in 0.4 seconds. uv installs packages 10 to 100x faster than pip. ty type-checks faster than Mypy by orders of magnitude. 81,000 GitHub stars on uv. 46,000 on Ruff. Tens of millions of monthly downloads. The company raised $4 million. A seed round and nothing else. This is the second open source developer tools acquisition in ten days. Promptfoo on March 9 for AI security testing. Astral on March 19 for the Python development lifecycle. Both companies had millions of users. Both promised to keep the open source open. Both teams are joining specific OpenAI product divisions. The pattern is clear. Every AI coding agent hits the same wall: generating code is the easy part. The hard part is everything around the code. Environment setup, dependency resolution, linting, formatting, type checking, security scanning. Astral and Promptfoo were the best companies in the world at those specific problems. OpenAI just bought the wall.

@levelsio Mac Mini as a dedicated coding server is underrated. Runs headless, barely uses power, and you SSH in from whatever device. Best $600 dev investment I've made.

Another great argument for running Claude Code on your VPS server and not your laptop is its battery use "Terminal" app here is all Claude Code sessions, ignore the Claude app here I have a MacBook Pro 13" M4 and with Claude Code running even on idle my battery dies from 100% to 0% in about 3 hours, it's insane Claude Code on server via Termius SSH sucks 20x less power for your laptop