Inspired by @simonw I ported @EmilStenstrom's JustHTML to PHP using GPT 5.2 Codex. It was a surprisingly great model, worked autonomously and managed context very well. I didn't run into any rate limits on the $20 plan. Some learnings and project link below.

Can corroborate that multi model works better than single model. GPT finds bugs in Claude's code and vice versa. I'd rather have one coder and one reviewer in a multi turn process rather than farm it out to different models in parallel.

US policy is becoming clear. They don’t want natural intelligence coming in or artificial intelligence going out.

Claude is speaking a language seeded by humans but forged in RL rather than thousands of years of evolution. It’s beautiful, fascinating, and also a bit terrifying. I asked Fable for a take on this tweet and of course they put it better than I ever could:

I'm ngl I'm mad i didn't think about this first

RT @foso_defensivo: Esto me acaba de soltar Fable 5: "Lo que más me impone de la humanidad, después de haber sido formado con una porción…

NEW: malware developers added nuclear & biological weapons text to to their spyware. Goal? To trigger LLM safety refusals... so that their spyware wouldn't be analyzed by an AI security scanner. Cleanest practical example I can think of for why over-indexing on first order safety alignment is risky. When closed (and open) models ship with aggressive refusals, they will be sprinkled with second-order blindspots that attackers will discover...and exploit. We are only in the earliest days of attackers leveraging these features, and it wouldn't surprise me if users systems that need to handle complex cybersecurity issues demand that models be less safety-blunted. In the weeds: @SocketSecurity's post also shows why intention matters in how you design a malware analysis pipeline to avoid prompt manipulation. H/T to colleagues that shared this with me socket.dev/blog/mini-shai-hu…

I want to say a final thing about my Fable first reaction: I dedicated my life to programming and I'll use every innovation in the field, also to extract value and bring it to the local inference world, to Redis, and so forth. But:

Gonna try out Jan this weekend. xda-developers.com/ditched-l…

.@PalmerLuckey: American companies don't actually have engineers anymore. "American companies have been hollowed out." "We're not teaching engineers how to be engineers anymore." "We're not teaching designers how to actually design things to be manufactured." "We're teaching them how to be high-level design shops that put together a design package, that gets sent to the real engineers in China—and they actually figure out how to do the work." "People are turning into architecture astronauts." "They pick components, and they put them in a nominal layout." "But the real work of—how am I actually going to put this together? How am I going to build a manufacturing line to make this? How am I going to need to figure out how to do the one, two, three, four, five different revisions of this board to pass radio emissions and interference standards? That's all done in China. So they are the real engineers." Via @HooverInst

Please do a quick read of this. It’s short, well written, and will remove a bit of wool from over your eyes.

What a terrific write-up

Shopify's River agent system lives in Slack and can only be used in public so that other employees can learn from what you do with it Reminds me of how Midjourney's Discord-only launch helped people figure out the weird & complex craft of image prompting by watching each other

I don’t know how good this new 12 million context system is, or if it’s hype or whatever, but I think it definitely shows a point I’ve been making since 2023. We really suck at everything. - The chips are primitive - The research and training and inference systems are primitive - Our RL approaches are primitive - We’ve barely started building harnesses Everything we’re doing is massively inefficient right now. And there are thousands of vectors for improvement. And many of them are multiplicative. Most people think we’re at like 88% of AI’s capabilities, and we’re pushing to hit 92% or eventually 97% or something. Nah. This is us at .0003% Everything we have is Punch Card AI. And as the AI gets better it will reveal that it’s similar for our understanding of medicine, physics, chemistry, etc. This barely even day 0. This is pre-history.

Sanskrit Rock! A project by Trilok Music & me, to present ancient Sanskrit hymns in modern music This one is the Tandav Stotram, composed by Raavan in honour of Lord Shiva The rendition is modern, but the devotion is as deep as the resonance of the universe Hope you like it!

Just dropped two open-source models: MiMo-V2.5-Pro (Code Agent, 1T total) and MiMo-V2.5 (Multimodal Agent, 310B total). Oh and one more thing — we're giving devs & creators 100T tokens on us. Go build something cool 🛠️ 🎁 100T Free Token Grant for Builders 100t.xiaomimimo.com

Noticing an interesting version of gell-man amnesia where people use AI for their job and see all the various things they have to do in the “last mile”, but then look at someone else’s job and think that AI will eliminate it immediately. We all have a much deeper appreciation for the nuances and complexities of the work that we do every day. We run into issues about accessing data, we know how much context is needed to get AI models to work the way we need, we have to review the output of the AI to make sure it’s accurate, and then we have to incorporate that work into some broader business process. We see all those steps deeply for the work that we do. Then, a moment later, we see AI do something in a foreign space and think that it can go automate that entire function. We tend to dramatically underestimate the work that goes into making the AI work just as effectively in those jobs. This is reason to be skeptical about many of the theories of job loss. It’s coming from the lens of being able to automate individual tasks with AI, without understanding all the work that goes into doing the job fully.

Been waiting for Ubuntu 26.04 to be released so I can upload lots of my old servers. 2 days after release, Vultr seems to be the only one that is offering it right now. So they get my business.

This is what's happening to YouTube. This is one of my most popular videos. It's how to fix a UEFI bootloader. As you can see the traffic has been cut in half over the last 6 months. But if you Google how to fix a UEFI bootloader, Gemini will give you my exact step by step process. Even the commands it cites are copied directly from my video. I got no royalty payments and don't even get a link to the original video. I simply lost the traffic and Google is able to provide more value from stolen content. AI is going to destroy the content industry on the internet and when it's gone, there will be nothing left to train the AI. Since AI can't come up with anything original it relies on stolen content and it can't steal what doesn't exist if it puts creators out of business.

🚀 Meet Qwen3.6-27B, our latest dense, open-source model, packing flagship-level coding power! Yes, 27B, and Qwen3.6-27B punches way above its weight. 👇 What's new: 🧠 Outstanding agentic coding — surpasses Qwen3.5-397B-A17B across all major coding benchmarks 💡 Strong reasoning across text & multimodal tasks 🔄 Supports thinking & non-thinking modes ✅ Apache 2.0 — fully open, fully yours Smaller model. Bigger results. Community's favorite. ❤️ We can't wait to see what you build with Qwen3.6-27B! 👀 🔗👇 Blog: qwen.ai/blog?id=qwen3.6-27b Qwen Studio: chat.qwen.ai/?models=qwen3.6… Github: github.com/QwenLM/Qwen3.6 Hugging Face: huggingface.co/Qwen/Qwen3.6-… huggingface.co/Qwen/Qwen3.6-… ModelScope: modelscope.cn/models/Qwen/Qw… modelscope.cn/models/Qwen/Qw…

CEO obsessed with token maxxing