Viewing profile — fooker

> you think kids in other states Parents of the kids, realistically.

We are not getting that reckoning. It’s great to yearn for bug free software, but that necessarily brings an insane amount of red tape to get anything done. Like 5 years worth of r…

> reword my initial prompt to get the agent off an unintended track. The signal here is the action of stopping the agent “do something else, this is stupid”, not a tweak to the ini…

Okay yeah, fair point. My comment was from a decade old perspective

Google has close to the best internal tooling in the industry for a decade or so. Then the Google engineers who joined Facebook missed it so much that they built a better replaceme…

There's a full fledged 'reasoning' step that basically expands your prompt. As long as you are not missing important information, how you word the prompt does not have any effect.

There's no program you run to 'make the thing'. It's mostly ad hoc scripts and some pretty horrible hacks being run by a hundred engineers trying to improve a thousand different th…

This trope was valid maybe in 2022. Model training now is not a straight forwards process of input data -> run tools -> get model. There's a whole lot of alchemy going on. We don't…

Exact prompts haven't mattered for about a year now.

Exactly right. Not only is it specialized, it also has its own field specific memes* and trends that evolve over time. The purpose is mostly to signal "I'm one of you". *memes as i…

I think, maybe you misunderstood my comment? I'm all for LLMs producing good research. That's obviously happening right now. What I said was people with good research being penaliz…

Publishing CS papers at the top venues requires using in-group language and formalism that is pretty much inaccessible to someone who has not done a PhD in that specific narrow fie…

How's that asymmettry worse than it is now?

Why would you assume that the same capability could not be used to harden systems?

Anthropic's fall from grace and mindshare seems rather accelerated. I wonder if these rapid movements are going to be the norm now. I imagine there would be angry investors if this…

Sure, agreed. The question then is, how much memorization is really needed for intelligence if you can query structured information?

I'm claiming that the max amount of intelligence you get out of a 100GB model is unlikely to be that much lower than what you can get out of a 1TB model.

No, it's like saying a chip that consumes 500W is going to be better than a chip that consumes 200W. This used to be true for the first several decades of microprocessors, and stop…

> Some groups are baking models into silicone While some other groups are baking silicone into models :)

Prediction - we are going to figure out SOTA AI performance without requiring 1TB of memory within a year or so. Of course CXMT, Micron, and family will still be profitable, but ma…

Great, so the other member of the set matters for you more than cost. Do you actually need to run the state of art model at 5 tokens per second instead of a qwen or whatever 7b or …

> Even if the output is like 5-6 tok/s, that might be usable for some purposes. You'll spend ~100x more on electricity than the API cost to have it run on someone else's GPU at sev…

agentic flimble please.

It's a basic step to make a plumbis

Not that you have much reason to believe a stranger on the internet, but I have seen this in person. Not doing acrobatics like the marketing video of course, just doing mundane thi…