High Signal Podcasts Evidence ledger
Method
Browse
← Back to evidence

Evidence receipt / belief

Published · transcript-backed

Joscha Bach: belief

1 Aug 2023 Lex Fridman Podcast #392 – Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans

“Now I think that almost every serious computer scientist goes through a phase like this in their life. This question that is almost no longer relevant because what is different between the programming language is not what they let the computer do, but what they let you think about what the computer should be doing.”

— Joscha Bach

Source trail

Everything needed to verify it.

Speaker
Joscha Bach
Attribution
Verified speaker
Claim type
belief
Recorded
1 Aug 2023
Publisher
Lex Fridman Podcast

Transcript context

…But it feels like it’s a tireless intern or maybe it’s an army of interns. So you get to command these slightly incompetent creatures and there’s an aspect because of how rapidly you can iterate with it. It’s also part of the brainstorming, part of the inspiration for your own thinking. You get to interact with the thing. I mean, when I’m programming or doing any generational GPT, it’s somehow is a catalyst for your own thinking. In a way, that I think an intern might not be. Yeah, it gets really interesting I find as when you turn it into a multi-agent system. For instance, you can get the system to generate a dialogue between a patient and a doctor very easily. But what’s more interesting is you have one instance of ChatGPT that is the patient and you tell it in the prompt what complicated syndrome it has. The other one is a therapist who doesn’t know anything about this patient, and you just have these two instances battling it out and observe the psychiatrist or a psychologist trying to analyze the patient and trying to figure out what’s wrong with the patient. If you try to take away large problem, for instance, how to build a company and you turn this into lots and lots of sub-problems, then often you can get to a level where the language model is able to solve this. What I also found interesting is based on the observation that ChatGPT is pretty good at translating between programing languages, but sometimes there’s difficulty to write very long coherent algorithms that you need to write them as human author. Why not design a language that is suitable for this? Some kind of pseudocode that is more relaxed than Python. That allows you to sometimes specify a problem vaguely in human terms and let ChatGPT take care of the rest. You can use ChatGPT to develop that syntax for it and develop new programming paradigms in this way. We very soon get to the point where this age-old question for us computer scientists, what is the best programing language, and can we write a better programing language? Now I think that almost every serious computer scientist goes through a phase like this in their life. This question that is almost no longer relevant because what is different between the programming language is not what they let the computer do, but what they let you think about what the computer should be doing. Now the ChatGPT becomes an interface to this in which you can specify in many, many ways what the computer should be doing and ChatGPT or some other language model or combination of system is going to take care of the rest. Allow you expand the realm of thought you’re allowed to have when interacting with the computer. It sounds to me like you’re saying there’s basically no limitations. Your intuition says to what larger language-…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.

Search evidence