Watch ChatGPT Deep Research do high level archeology, product development, and algorithms

[@dwiii1635]

How you guys only have 848 subscribers is beyond me. Good shit.

[@TheFeatureCrew]

Appreciate it - on that grind to 1k 🙏🏻 - Jakob

[@georgemontgomery1892]

Ugh, I am finally subscribed here. I see these videos and love them but always forget to sub until now. Great work guys.

[@TheFeatureCrew]

Thanks for the support, George! - Jakob

[@vassovas]

Have you guys tested Google Deep Research? Should do comparison episode

[@TheFeatureCrew]

Good call, putting it on the list - Jakob

[@adolphgracius9996]

I have used Deep Research from Google and I can confidently say that this is at least marginally better in about every metric that matters

[@CosmicCells]

I also have used Gemini Deep research and from what I have seen and heard on KZbin (I dont have Pro sub for OpenAI), pretty much everyone agrees that OpenAIs version is a clearly noticable higher quality output. It usually uses less sources but the outputs but goes much more in depth. I think its cause OpenAI is using a very powerful reasoning model in the background while Googles Gemini 1.5 pro is an older non-reasoning model.

[@sethjchandler]

I wish you would test it on legal analysis.

[@JohnGodwin777]

Good work, thanks

[@TheFeatureCrew]

Thanks for watching! - Chris

[@ListenGrasshopper]

Have it do this..."Analyze the potential long-term economic and social implications of artificial intelligence on global labor markets, comparing perspectives from three different academic disciplines: economics, sociology, and computer science. Construct an argument that weighs both the transformative potential and the significant challenges, using specific case studies from at least two different industries (such as healthcare and manufacturing). Critically evaluate the potential for job displacement versus job creation, and propose a nuanced framework for workforce adaptation that balances technological innovation with human capital development."

[@augmentos]

is the prompt and prompt techniques the same and as important as regular models even 03-mini-high or PRO? Or is it better to be vague so it can go where it wants? Loved this group discussion. Subd

[@HasanFaqir]

Great deep dive! At 0:01 and onwards the way you zoom in on the two people it’s cropping both of them and it’s kinda annoying.

[@OddlyTugs]

Compare with the open source deep research things. Try it with some local models

[@TheFeatureCrew]

Planning on doing this - have any favorite os researchers? - Jakob

[@OddlyTugs]

@@TheFeatureCrew smolagents looks promising

[@joeincognito8940]

@stefanmilo