74 karma · joined March 24, 2014
Down the line, I think models will start using video generation as part of how they “think.” Picture a version of GPT that works frame by frame—ask it to solve a geometry problem, and it generates a sequence of images to visualize the solution before responding. YouTube’s massive library of visual content could make something like that possible.
Think Amazon that has both AWS and the retail business. There's a lot of value in providing both.
Those are words Trump has used. He said the eating pets thing during the debate.
https://www.reuters.com/world/us/inside-trumps-plan-mass-dep...
What is that? Supply chains have improved. The labor force has expanded, partly due to increased immigration, and that's helped to take some of the edge off of the supply-and-demand imbalances that we had when inflation was very high two years ago." https://www.pbs.org/newshour/show/examining-how-economic-pla...
Immigration: "After hitting a record high in December 2023, the numbers of migrants crossing the border has plummeted since then. Harris and the administration have credited their tough anti-asylum measures for stemming the flow, although increased enforcement on the Mexican side has also played a key role." https://www.ap.org/news-highlights/elections/2024/where-trum...
Abortion rights: "At one of her first campaign events, she stated that if Congress “passes a law to restore reproductive freedom, as president of the United States I will sign it into law.”" https://www.aclu.org/news/reproductive-freedom/how-kamala-ha...
If you don't like what her positions are that's your prerogative but it's just not true that she did not have answers to these questions.
- write me a file with the function definitions for this problem. - compile that - write a test that test x outcome - compile that - then have it start writing functionality
If it's trying to one shot a complex problem that you would typically break up, your prompt is probably too vague.
It seems like you used an interpreter in the loop which is likely to help. I'd also be interested to see how o1 would do in a task like this or if it even makes sense to use something like prolog if the models can backtrack during the "thinking" phase
There's also Unacademy a company I'm sure has helped millions of people learn new things.
I did try to Google it first, but wasn't able to find anything on it.
And we actually do have alerts that go to the campaign team if a voter asks something that the agent can't answer.
I don't see how it's anymore propaganda than having them hire a bunch of people to text you.
My view here is a bit more optimistic. LLM will allow campaigns where each individual essentially has a personal outreach plan. It's like a candidate can have a personal conversation with every voter. This will make things less divisive because you can talk to each voter about things that matter to them. There's no need to say the most inflaming thing just to get engagement.
One other bonus is that smaller campaigns can get their message out in a way that's authentic to them. If you're a parent running for school board because you care about your child's education you don't have an army of volunteers to do the outreach for you.
What do you think about the tradeoff of having an agent manage its own memory as opposed to having a separate agent whose job it is to manage the memory and the other agent just focuses on the conversation?
I do think this will be way less than having all of the functions listed to begin with though. I think the discoverability is a novel approach. Honestly, I'm surprised ChatGPT with plugins doesn't do something like this by default rather than making you pick which plugins you want at the beginning of the conversation.
I am curious about the benefit of having the agent interact with the user (or doing the task) and managing its memory instead of having an observer agent that modifies the memory separately. The thought process is to let the agent use all of its tokens to focus on the task and not memory management.
You mention that a language model can't model language, only text, because it only observes examples of text, not language. However, I would argue that these text examples are in fact representations of human language ability. The text that language models are trained on is a product of human language ability. This text, while not a perfect representation of language, is a direct outcome of language use and thus carries within it the patterns and structures that language models can learn.
Moreover, models like Transformers do have a sort of latent space - the embeddings space. This is a continuous space where words and phrases are represented as vectors. The distances and directions between vectors in this space capture semantic and syntactic relationships, which suggests that these models are indeed learning some aspects of language, not just text.
On top of this, many of these models are trained on diverse data sources, including transcriptions of spoken conversations. This introduces elements of pragmatics, or how language is used in real-world conversations, into the training data.
Finally, the translation capabilities of these models further suggest that they are capturing something beyond mere text. They are able to translate between different languages, which implies that they are learning some underlying linguistic structures that are shared across languages.
Again, it's important to stress that these models are far from capturing the full complexity of human language ability. However, to say that they only model text and not language seems to me an oversimplification. They are learning patterns and structures in the data that are intrinsically tied to human language use, and so in a sense, they are modeling aspects of language.