Dagens Vibes — 3. august 2026

Qwen@Alibaba_Qwen

Meet Qwen3.8-Max: A New Bar for Coding and Cowork.

Andrej Karpathy@karpathy

We're starting to leave the territory where you'd test an LLM by e.g. "create an svg of pelican on a bicycle". As one idea to generalize it, I was interested what Opus 5 would do if I gave it the first paragraph of the Lord of the Rings, a 1M token budget (~$10) and asked for three js render of it. Opus went off for ~2 hours and wrote 5500 lines of code that (procedurally) rendered the story. It's kind of janky but fun. But it's a bit mindboggling that the LLM has to place and orchestrate various polygon assets in (x,y,z) coordinates and write code that animates it all, and that it even does anything at all. I also like this kind of examples because no one in their right mind would ever spend the time to write something this custom but LLMs have all the stamina and patience in the world, so it's an example where we go from "no one would ever do this" to "sure, why not, it's ~free". There might be a lot more. But I'm excited about creating hyper custom worlds that you can imagine dropping players into, e.g. here to participate in the LoTR story as a spectator NPC, or one of the characters, or etc. Something like an ephemeral GTA of X on demand. Last thought is that the domain of worlds/games exposes a weakness in LLMs: they can't easily audit their work because they aren't able to efficiently and natively perceive videos or play games within them. Here, Opus 5 had to very slowly and painstakingly take screenshots at different points, and it messed up a few times and created a bunch of jank. An example of raw capability (multimodal, gameplay) that I think is still quite lacking.

David Crawshaw@davidcrawshaw

Machines create code now. That means machines can personalize your software, if it is open source. Closed source tools (like Claude Code) are now less desirable. https://blog.exe.dev/devtools-must-be-open-source

Machina@EXM7777

listen to me... ALL YOU NEED is screen time like literally, just spend your days inside agent harnesses building projects from scratch: - learn how to build your own harness (pi / omp) - set up your own agent configs, change the rules, watch the behavior shift - steal skills from public repos and bend them to your workflows - feel the exact moment context starts hurting your output instead of helping - build your own knowledge base and push it til retrieval breaks - learn what effort level each task actually needs knowing how agents work under the hood gives you a clean start... but the real edge is a vibe that only comes from reps right tools, right config, a real feel for orchestrating harnesses and you can achieve basically anything get your screen time up

Oren Melamed@OrenMe

The integrated browser in @code is getting another interesting upgrade - you can now comment directly on elements in the page, with numbered pins, previews, and editing. Those annotations are synchronized with the chat input as attachments. So instead of: “Hey, change that button on the page” You can point at the exact element, annotate it, and hand that context directly to the agent. This is also more interesting cause it makes the browser a great tool for product and UX iteration

Ben Davis@davis7

Add this to ur Codex config, makes it so that Codex will use the multiple choice tool outside plan mode. Makes it feel 100x better