quicklywilliam

I think the big idea here is that you can get a lot more performance if you take an integrated approach. This specific model made to work with this specific inference engine made to work with this specific harness/agent. When everything is done separately, developers of a given pieces have no idea what they are targeting for all the other pieces.

This is currently a huge advantage that Anthropic has over open weights models – they control the whole stack. Indeed, they train new models against Claude Code!

It's early days on this project, but just imagine it gets enough traction that future models start training against ds4. Indeed, in the post Antirez even seems to be hinting at some sort of collaboration with DeepSeek?

Picture the early days of the PC - local models are like that. While there are scattered groups of hobbyists who tinker endlessly to get a few lights blinking, almost everyone doing “real work” is using a large, expensive, centralized product. Despite the exponential performance curve of smaller models (not to mention the rapidly increasing cost of using frontier models), we are all still stuck using mainframes and minicomputers. Why? Because few of us have the time and resources to build a computer from raw components.

The problem with local models isn't that they are local, it's that they are not integrated. Hobbyist testing the latest local models are great (thank you!), but we also need someone building the Apple I.

Same question. Given the people involved I am inclined to take it seriously, given the money and hype involved I am somewhat less inclined.

Their website doesn’t even have a hint of what the approach is.

Agreed, there is probably a theoretical world where we got enough money/compute together and had this explosion happen earlier.

Or perhaps a world where it happened later. I think a big part of what enabled the AI boom was the concentration of money and compute around the crypto boom.

This is a really cool list and repository of ideas. Seems like the focus of the work is on making knowledge legible to AI. I wonder if you (or others) have done a similar level thinking about the inverse – making AI more legible to humans?

Curious if you looked at using SwiftDown (https://github.com/qeude/SwiftDown), MarkupEditor (https://github.com/stevengharris/MarkupEditor) or any other libraries for live/WYSIWYG Markdown editing.

There’s been a lot of discussion lately about Anthropic and others turning the screws on their subscription plans, skyrocketing costs for enterprise customers. This is driving more and more folks to consider cheaper and non-proprietary models (which are getting more capable) for some of their tasks. This article goes into both of these trends.

I wrote a thing about why the current direction of AI tends to produce such terrible work — and what I think we could be doing with AI instead.

The real question is who bought the other 5,742 of them

Interesting read. I don't know if I quite buy the evidence, but it's definitely enough to warrant further investigation. It also matches up with my personal experience, which is that tools like Claude Code are burning through more and more tokens as we push them to do bigger and bigger work. But we all know the frontier model companies are burning through money in an unsustainable race to get you and your company hooked on their tools.

So: I buy that the cost of frontier performance is going up exponentially, but that doesn't mean there is a fundamental link. We also know that benchmark performance of much smaller/cheaper models has been increasing (as far as I know METR only looks at frontier models), so that makes me wonder if the exponential cost/time horizon relationship is only for the frontier models.

Great idea and implementation! If you are hesitant to install this for any reason, you can accomplish the same thing with this one liner:

  sudo bioutil -ws -u 0; sleep 1; sudo bioutil -ws -u 1

There's been a ton of debate about whether we are in an AI bubble, but I've read a lot less about what will happen when the bubble bursts. I took it for granted that this is a bubble, and in particular that frontier AI companies are heavily distorting the market by subsidizing inference costs. Here's what might happen if those subsidies suddenly go away.