cloudking

For ChatGPT I've found this search extension useful to find previously used prompts: https://chromewebstore.google.com/detail/gpt-search-chat-his...

Source code: https://github.com/polywock/gpt-search

It works but what are the next steps for checkpoints compatible with Apple chips?

LLM?

AI Jeeves would be killing it right now

As a term? No. But I've heard people saying "ask ChatGPT"

It's 100% local only. No server side component, analytics or ads.

There's a bunch of commercial options live right now that have comparable results. Not sure what all the hype is about, here are a few:

https://runwayml.com/

https://klingai.com/

https://hailuoai.video/

https://lumalabs.ai/

https://pika.art/

https://viggle.ai/

Congrats on the launch, this is an interesting concept. It's somewhat akin to developers approving LLM generated code changes and pull requests. I feel much more comfortable with senior developers approving AI changes to our codebase, then letting loose an autonomous agent with no human oversight.

Translation: here's a bunch of useful features requests customers asked for, but we don't have engineering budget to build them and not enough customers asked for them, so we're not building them. Plus they don't say "AI" so our executive team said to axe them.

Super hot

Good question. Chrome itself isn't a standalone business, the money generated through Chrome still primarily comes from Ads. The hardware tied to Chromebooks generates some revenue, but even ChromeOS is essentially free. They generate a tiny amount of revenue selling ChromeOS management tools in Workspace. Why not spin off an actual revenue driver like YouTube?

So far from what I can gather, the concept of "AI employees" or "agents" are simply a series of prompts and functions daisy-chained to perform tasks. There's not much fundamental difference between an "agent" and an "app", other than there being some AI models under the hood. You still need a human to drive the agent, they aren't off performing tasks without being instructed or configured to do so. The current paradigm very much complements human workers who are leads, generalists, managers etc and technical enough to know how to drive them. So I don't think these higher level jobs are impacted negatively as much immediately with the current tech, if anything they are slightly more productive (note this is Nov 2024, I'm not saying this can't change, but this is the current state).

Will the entry-level jobs be impacted? Most likely they are already, anecdotally as a generalist I can perform a lot more tasks than I could a few years ago because of language models. I no longer need to hire junior developers to create prototypes and scripts, I can create them myself. Similarly, I don't need to hire copywriters to create documentation or marketing material. I've also started leveraging image generation models for basic things like landing pages, so I don't need photographers or stock image sites anymore for those tasks. I have meeting assistant "agents" taking notes and action items, I don't need an admin assistant at meetings. There is definitely a shift happening at this level, right now.

main.24cd0805.js:1 Uncaught SyntaxError: Unexpected token '<'

Innovator's Dilemma

I've heard Meta's Demucs is SOTA, has anything else better come out since?

Wonderful project. Built video games and music visualizers that I never thought were possible on web.

I granted mic permission, I see my mic input working, but the avatar is not responding

This seems interesting, how do you actually decide what to fetch from the memory that is relevant? "This setup enables efficient memory retrieval based on relevance, recency, and context" -- but how does it actually "decide" what to retrieve?

Nice! Take a look at LiveKit they have a Python implementation too https://docs.livekit.io/agents/openai/multimodalagent/

If you really care about this kind of stuff, a simple AdGuard or Pi-hole setup can block all these requests across your network.

It seems to come naturally when the AIs solve hard problems, I end up thanking them out of habit.

Palantir has already been using these models since they came out, you can see Llama 2 models in the second screenshot https://www.palantir.com/docs/foundry/functions/language-mod...

Very well done! Maybe add a bonus multiplier for lower completion time

M3 Pro user here, agree with the same points. It's nice to be able to actually have your laptop on your lap, without burning your legs.

I think the value here is not in searching for SEO crap, but turning it on when you want to get references to the most current information relevant to your query.

For example, if you ask LLMs to build code using the three.js library, nearly all of them will reference version r128. Presumably because that version has the largest representation in the training data set. Now, you can turn this on and ask it to reference the latest version, and it will search the web and find r170 and the latest documentation to consider in it's response.

I was already doing this before by adding "search the web for the latest version first" in my prompts, now I can just click a button. That's useful.

Makes you wonder who is sponsoring these articles.

Are you using the @codebase feature? It will perform a search across the codebase for files relevant to your query and suggest changes across files.

https://docs.cursor.com/chat/codebase

Necessary because a user can always inspect > modify the HTML.

What is AGI and ASI? I think a fundamental issue here is both are sci-fi concepts without a clear agreement on the definitions. Each company claiming to work towards "AGI" has their own definition.

How will someone claim they've achieved either, if we can't agree on the definitions?

OP is demonstrating that the product follows prompts from the pages it visits, not just from it's owner in the UI that controls it.

To be fair, this is a beta product and is likely ridden with bugs. I think OP is trying to make a point that LLM powered applications can be potentially tricked into behaving in ways that are unintended, and the "bug fixes" may be a constant catch up game for developers fighting an infinite pool of edge cases.