Perhaps we'll get to a point where believing any un-sourced information from an LLM will feel crazy. I don't want my model to know more than it needs to perform logic and use tools. Once it is capable of using tools I would much rather it looked up information or sourced it from existing context rather than just divine it from it's weights.

The problem is it needs world knowledge to know what to lookup. This puts a floor on how little it can know while being able to look up what it doesn't know. Maybe its better if it knows a lot but has a good instinct for verifying that.

How do you actually get to the thing you're looking up? The scrapeable internet is shrinking in response to scrapers.

"Source: rare book ingested and shredded by Anthropic. No, you can't look it up and we can't show you the scan. The remaining open market copy is $5000. Trust me."

I empathize, and I have the same preference, but I wonder how this interacts with other people (many of them being our coworkers) using LLMs. There is no authoritative source for the models to pull info from, so either people will have to exercise good judgement and double check important claims, or they will trust too blindly and fall close to the level of whatever LLM they use. In that case, I prefer my coworkers to use an LLM that does have world-knowledge -- I will still hear them spout ridiculous claims, but at least it should be less frequent. It strikes me there's a sort of prisoners dilemma here, where if nobody trusts others to critically evaluate info, it's in our interest to make the tooling do it instead, to whatever degree that is possible. Maybe I'm too cynical about working with others though.

Really want to play this with a controller. Seems to me that a joystick to steer and another to adjust the sail would be really intuitive. Aside from that really fun, looking forward to future updates.

Thank you, yes! We’re planning a steam release soon where we would add controller support on launch

Except it is intentionally unstructured. AT proto is one way to add the structures we obviously care about for "social" internet media on top of the freedom that the web as a whole provides.

Eerily close to Roman decimation.

Where do you think Jack Welch got the idea from?

Came here to say this, difftastic is great as long as you are working in a language where it understands the language tree (most languages). Getting away from diffs being focused on line changes to diffs that understand the actual language makes so much sense once you start to use it.

Ideological is carrying a lot of weight there. Perhaps you can be more specific about the ideological arguments you are hearing that are not worth it?

Walled gardens have less fraud and malware because it's less open. But developers prefer open source decentralized software. Of course, we are technologically literate enough to avoid the fraud. It's similar to drug decriminalization or the legalization of sports gambling.

Yes as of about a week ago, last I checked.

Anthropic is actively blocking calls from anything but claude code for it's claude plans. At this point you either need to be taking part in the cat and mouse game to make that plan work with opencode or you need to be paying the much more expensive API prices.

i see

i guess they were blocking OpenCode for a reason

this will put people to the test that use mainly Anthropic, to have a second look at the results from other models

I'm not particularly against AI programming but I don't think these two things are equivilent. A compiler translates code to specifications in a deterministic way, the same compiler produces the same output from the same code, it is all completely controlled. AI is not at all deterministic, temperature is built into LLMs and furthermore the lack of specificity in prompts and our spoken languages. The difference in control is significant enough to me not to put compilers and AI coding agents into the same catagory even though they are both taking some text and producing some other text/machine code.

Existing social platforms are built for profit, which modulates emotion for engagement (something kind of like entertainment, but I wouldn't say I'm entertained exactly by the rage bait I'm often fed by algorithms). Users of an open protocol might select for the same experience, or they might not, I think that's yet to be seen. This also assumes that this fantasy open protocol could also escape the pressures of maximizing profit.