I think that in the process of actually struggling to craft a game people end up just doing lots of playtesting and at least if someone produced something as polished (visually and technically) as these are it'd be somewhat fun at least for the person that made it. As in it used to be very unlikely to get something that is technically complete and visually polished, but not game-play polished.
> I turned it on briefly to receive a confirmation code that I needed in order to maintain that phone plan itself, and once spiritually but not literally - when going on a remote and slightly dangerous hike, friends insisted that I bring along a few communication devices that they had for emergency comms and use them to check in once every few hours.
It is funny to have witnessed the conversion from the default state of people is being not connected digitally to being connected digitally and once people start living with the 24/h connection they start the perceive simply being out of cell network as carrying inherent risk.
I've noticed people reacting to going hiking beyond cell networks with a "but that's dangerous!" response. Being outside of cell networks was the natural state of being for the vast majority of human existence!
> I've noticed people reacting to going hiking beyond cell networks with a "but that's dangerous!" response. Being outside of cell networks was the natural state of being for the vast majority of human existence!
And it was dangerous.
Also: I grew up just before cellphones started to be a thing and I felt a lot of times as a kid the frustration of not being able to contact someone remotely.
Which specific model? and are tools available? I would say that there is lots of RL with tool use, but that the reasoning traces are not always perfect. In general, models think about things that might be wrong, but will recover from this poor line of thinking.
This was using llama-serve on my local which by default gave it current time & unhelpfully "get_info" which is specific to the browser
The full request was
```
{"messages":[{"content":"What animal lays the biggest egg?","role":"user"}],"return_progress":true,"sse_ping_interval":1,"stream":true,"tools":[{"function":{"description":"Returns the current local date and time in ISO 8601 format, with the IANA time zone name","name":"get_datetime","parameters":{"properties":{},"required":[],"type":"object"}},"type":"function"},{"function":{"description":"Get runtime info (OS name), may call when user asks about local files or shell commands","name":"get_info","parameters":{"properties":{},"required":[],"type":"object"}},"type":"function"}],"reasoning_format":"auto","reasoning_control":true,"backend_sampling":false,"timings_per_token":true}
```
Disabling tools removes the bit about mentioning instructions.
Just an interesting detail I noticed.
Do y'all have any penalty/regularization w.r.t. reasoning traces or is it pure SFT -> RL?
interesting, I didn't see it on the playground or on my local llama-server.
SFT -> RL. In our fine-tuning notebooks, we typically say that rollouts that go past the max context length (too much reasoning) get 0 reward, so there is slight pressure for shorter reasoning, but I don't think we have strong regularization. I think the agreed-upon recipe is to first make the model smarter (longer reasoning), then teach it to reason faster/denser.
(I work at Liquid.) Probably, yes. Per the blog, the last 3.5T tokens of pre-training were a mid-training stage with a larger share of agentic traces and tool calls, and SFT covered tool calling and agentic traces too, because the text model is post-trained to live inside agent harnesses. With an empty system prompt it seems to assume it's in one. I'd try a short system prompt that says no tools are available, and if it still does it, a note on the Hugging Face model page would help the team.
Do you use API for google & anthropic or subscription? I was under the impression those two are apt to ban you for using their models in a non-approved way
I remember fighting with business about this in a startup way before LLMs, more code, velocity or motion doesn't mean more value.
If I put on my optimist hat for a moment, I hope LLM's would finally make this obvious for people, more stuff !== more value. 250k lines of code a week isn't a flex IMO, it doesn't really mean anything out of context.
There are a few one-shotted personal single page apps, yes. Mostly useless beyond an initial "that's cool" factor. Pre-LLM that was occupied by the "I hacked this in a weekend" niche.
reply