297 karma · joined April 14, 2022
Maybe there should be. Interesting idea.
I feel like we should have all moved on from the fact that LLMs don't have a very precise memory? This is (not entirely, but mostly) fixed by including the information you want to query in the prompt itself, like Bing does. ChatGPT is much better at summarization than precise fact recall.
And most of the time, it doesn't matter. When I'm using ChatGPT to learn or troubleshoot something, it's a jumping off point, not the end of the story. As long as everyone understands that, there's no problem. I have yet to see anyone who treats its output as 100% fact, and no AI company claims it either.
Would not be surprised to find this fear was astroturfed. It makes literally no sense.
But if you look at projects like Autogen ( https://github.com/microsoft/autogen ), you see one master agent coordinating other agents which have a narrower scope. But you have to create the prompts for the agents yourself.
This GPTs setup will crowd-source a ton of data on creating agents which serve a single task. Then they can train a model that's very good at creating other agents. Altman nods to this immediately after the GPTs feature is shown, repeating that OpenAI does not train on API usage.
Prediction: next year's dev day, which Altman hints will make today's conference look "quaint" by comparison, will basically be an app built around the autogen concept, but which you can spin up using very simple prompts to complete very complex tasks. Probably on top of a mixture of GPT 4 & 5.
https://thinkml.ai/iphone-ai-artificial-intelligence-feature...
I don't agree with a lot of his manifesto. But I wish more people would write manifestos and try to fully explain their worldview instead of spending a lifetime reacting to other people's opinions. It's sad that we think they're inherently pretentious.
What does disappointment me about this industry is that we haven't found or invented some low-code app builder that is actually good enough for most use cases. I feel like we really should be spending most of our time as logicians who encode business logic, but we spend way too much time churning through supporting technology.
But every app builder I've seen runs into the same problem, that complexity scales exponentially in complexity a way that code doesn't have to. So it quickly becomes better to have built the dang thing from scratch.
And it's not like I can think of anything better. So it feels like we're missing something fundamental, probably to do with how we imagine user interfaces. Hopefully LLMs open the door to something better here eventually, and I'm not at all saying that chat views or plain speech are the best user interfaces we can do.
At a certain point, apathy sets in. If this is our generation's mesothelioma, so be it.
Not ideal since people make packages as a labor of love. Maybe a community fund to pay gas fees would work.
This is not true - zero knowledge proofs allow you to verify work much more cheaply than the original computation, and they're used heavily in upcoming ethereum L2s.