binyu

It is much more of a context window size and model capabilities problem. Local models are not even remotely close in solving complex problems, even when used with the same harness.

That is true, it is a 1.6T parameters model so it requires a great deal of memory. I also heard there's a 2bit quantization that works well on Apple metal.

Right, thank you. Personally I think reading all the code that the AI produces is impossible and kind of defeats the purpose of using it. The key is to devise a structured way to interact with it (ski…

They all still fall short of Opus 4.6, definitely though. They are good but fail on extremely complex tasks, in contrast with a frontier model that will keep on trying until it succeeds or exhausts th…

DeepSeek V4 with 1 million token context window is pretty powerful, although still not there. There's hope that Opus 4.5 level performance locally is not that far away.

Gotcha, that implies you are going to read the code that the AI produces anyways. > Go reads fine whether the architecture is good or bad Were you reading the Golang code all along and got fooled o…

> I'm rewriting k10s in Rust. Not because Rust is better but, because it's the language I can steer. I've written enough of it to feel when something's wrong before I can articu…

Yeah, that part is probably not done by Claude.

I don't think this protects from sidechannel/timing attacks applied to images load time completely. Edit: Reading more thoroughly, probably it does to a great extent after all.

They forgot to add timing attack on images load time which can be used to tell if you visited X website. https://www.ieee-security.org/TC/SP2011/PAPERS/2011/paper01…

Hey, this is cool work. By any chance, did you see Cloudflare Artifacts? https://blog.cloudflare.com/artifacts-git-for-agents-beta/ …

Pinocchio had definitely some sociopathic traits as a character. The Pinocchio movie by Walt Disney from 1940 is a masterpiece. Highly recommended.

Very cool. Was this inspired by Cloudflare Artifacts? I'm working on a similar system as the backend for an upcoming agentic coding platform.

> which is easy, and more like changing the colour of a carbon-fibre Formula 1 part, which requires re-calculating the weight, strength and aerodynamics. Seems like a bit overstated for a CPU fan b…

This is really cool. It has retro vibes of the era when the Internet was still free from the big five domination.

> At least crypto does not take away more jobs than it creates Except sometimes when there's a huge black swan event, or when the bubble pops. Such things can result in significant layoffs eve…

How does this address the point I moved specifically?

I feel like Anthropic is forcing their new model (Opus 4.7) to do much less guess work when making architectural choices, instead it prefers to defer back decisions to the user. This is likely done t…

Yeah, people really took it to the extreme and made a cult out of it for no reason. Mass delusion at its finest.

Finding specimens is not that hard or inaccessible if you are determined. Virtually any place on earth has its own geomorphology history. Start by looking at geological maps to learn what kind of rock…

Interesting perspective. While the analogy may be somewhat intuitive, distributed computing exhibit a wider and more diverse set of challenges imo. Example: Synchronization in naturally async environm…

Frontier models being in the hands of a handful companies does not help either. Let's hope that the open weight movement changes that soon.

I disagree with you. Today's scenario is certainly much more interesting than the post dot-com boom years, maybe not as interesting as the very early days of computing, but certainly ripe for inn…