large language model

“I might be wrong, but I think that eventually we will find ourselves in the world of hybrid models and multi-models where there will be several LLM models coming together because each one of them will have their own benefit. And then there's going to be a leas…”

“Like, LLMs may make it possible to bring on a supply Type that maybe the long tail that was just, it was too much effort to reach out to them, onboard them, but maybe if you automate that work, you actually create an opportunity to expand the supply in a way t…”

“Again and again, what you see, and I think a lot of this for new companies kind of leveraging large language models to build products is that focus on a constrained market to start is the way to win.”

“What sometimes people will do is they go, oh, there's this really popular, large language model, whether that's a commercial or open source, and then they make everything fit into it. And what you end up doing is actually reducing the quality that someone can …”

“You don't need to create an LLM because it's a commodity thing now, and there's a bunch of providers who can do it way better and have way more resources to do it with than you do.”

“What I can do is take those transcripts and we literally did this. We fed them into some of these large language models that have super long context windows now. And what you can do is actually pull out what's the most common problems that come up, you know, a…”

“It's a data management problem. Like, it's getting access to good data, getting access to high quality data, getting access to timely data, and getting it to the LLM to get the LLM to make a smart decision. That's where 90% of the calories go.”

“What will make my AI experience really great? It's definitely not going to be the models because like these models are mostly going to be somewhat replaceable and you could say, okay, well, is it going to be the prompts? Maybe, but you know, so many good promp…”

“Usually I will put my additional information at the beginning of the prompt. And that is helpful for two reasons. One, it can get cached. So subsequent calls to the LM with that same context at the top of the prompt are cheaper because the model provider store…”

“If you end up vibe coding this prototype, but you start off with a very clear plan about this is what the thing needs to look at, look like rather than going immediately into a conversational mode where you're sort of co-designing with the LLM and going back a…”

“Fast forward to today, we're working with Seven of the frontier labs, basically every lab that's doing work and building the best large language models.”

“Companies where their code base remains and they are attaching AI to it often aren't the ones who are going to revolutionize their industry. Companies for whom their code base perhaps shrinks and the LLM becomes a core part of what it is that they use to solve…”

“There's this skill of being able to write evals. I know everybody can write prompts, prompt engineering. You can try and focus the LLM so that it can offer better insights and offer better results. But even as your LLM actually Provide, produces results. You n…”

“If you feed in derivatives of derivatives into the model, you will basically take the wisdom of the crowd and that will shrink and you'll have a single opinion on everything, which is really bad.”

“And I think also the LLM is, is probably not going to say your product if it didn't show up anywhere on the RAG. So I think that's where most of the interesting stuff is for, from an optimization perspective.”

“What we usually find when we try to ask an LLM to do this error analysis is it just says the trace looks good because it doesn't have the context needed to understand whether something might be, you know, bad product smell or, you know, not”

“The tricky piece here is that it's not enough to just wait for that technology to get better and then start building on top. You have to build beforehand to like make a bet and then that's the LLM to catch up because when that model releases, you already need …”

“You don't know how the user might behave with your product, and you also don't know how the LLM might respond to that, so you're now working with an input, output, and a process, and you don't understand all the three very well.”

“If you just go and you prompt and you prompt and you prompt and you prompt, you'll realize that no matter what tool you use, the memory just isn't infinite, right? By the time you reach message number 10, 15, 20, 3040, snippets of early messages sort of get lo…”

“The highest bandwidth API for an LLM is just chatting to it, is just saying like, I have Gmail and I have these folders that are really disorganized. And what I really want is to be able to come into my inbox every day and really know what's important, get rid…”

“There appears to be a very strong correlation between how good their drawing of a pelican riding a bicycle is and how good they are at everything else. And nobody can explain to me why that is. But as I started looking at these things, I realized, wow, The bet…”