LLM Models
topic on 7 shows · 12 statements across 10 episodes
BG2 Pod 
Innovators & Investors 
Latent Space 
Invest Like the Best 
the a16z Podcast 
Big Technology 
20VC
12 statements about LLM Models, every show
Krentsel: Decoupling executor from harness enables safe agent self-evolution
“What this gives you this nice split is it gives you an isolated execution environment down here. It gives you protected state, and then it has a very explicit stateless layer that's safe for self evolution. So the executor can propose changes to itself in a wa…”
Abrahami: Standard LLMs lack technical capability for astrophysics research
“You cannot do research on, on astrophysics with LM. They don't have the capacity for that. It's not what they do. You're gonna, they might answer you, but they don't run the models, the simulation, the background.”
Bayramov: Even engineers building LLMs cannot explain model behavior
“LLM models they It's pretty much like a black box. You don't know what happens to the data that goes in because like even the engineers who are building those models, they cannot explain the behavior of the model.”
Zhang: LLMs abstract rigid conversational decision trees into neural weights
“In the old days, to map out a conversation, you just build, like, a gigantic tree, right, of decisions, and that's very hard, because no one likes that experience, and, like, you ask something that's not quite one of the branches, and it just forces you down t…”
Dax Reed says complex optimizations on top of LLMs quickly become obsolete.
“Any kind of fancy optimization you're doing on top of the LM eventually gets, like, you kind of lose the need to do it given how fast things are moving.”
Appenzeller: Guaranteeing LLMs never produce forbidden output is an unsolvable problem
“You're trying to have an LM that is very helpful and never even implicitly gives investment advice. That's sort of an unsolvable problem, right? You can get better and better and better, but you can never completely rule it out. And you can add a second LM tha…”
Mlejnsky: Smarter LLMs reduce the need for prompt management tools
“Good question to ask when thinking all about dev tools with LM. Is my dev tool more relevant as the LMs are getting smarter, and as people need less prompting? I'm, for example, like, really bad at prompting. But like, I can get more work done over the years b…”
Petersen: LLMs parse changing spreadsheets into databases without rewriting parsing code
“It'll take a spreadsheet, which is like I was saying, is just like massively complex spreadsheet and put it into a database accurately, like more accurately than humans doing data entry. And these spreadsheets change every week, every few weeks. It's not, it's…”
Gurley: LLM scaling on parameters and context windows could hit ceilings
“I look at the LLM models. I mean, they've scaled up on parameter count and the width of the attention window. There easily could be limitations to that. Like if you just think about how optimization models work, they could run out. Like it's not infinite scali…”
Socher: Base LLMs will become commoditized like databases
“And I think it won't matter that much which database you use, just like it won't matter that much, which LM you use, but it matters what you do with it, how you tune it what kind of training data you add onto it to fine tune it.”
Socher: LLMs are inherently ill-suited for numerical financial forecasting
“Like an LM won't be as good in doing a financial forecast because that's not what they're trained on. Like LM's help a ton for all things, natural language, because they understand so much about natural language and will have so much world knowledge, but that …”