OpenAI’s ChatGPT o3 – Pushing Humanity Forward!

❤️ Check out Vast.ai and run DeepSeek or any AI project:

📝 OpenAI's o3 is available here:

📝 The paper "Humanity's Last Exam" is available here:

ChatGPT trick:
Go to your profile – personalization – customize ChatGPT, and add:
“Look for peer-reviewed sources for every answer. Rate the credibility of that source on a scale of 0 to 10 for every source. And of course – use more emoji. ”

📝 My paper on simulations that look almost like reality is available for free here:

Sources:

Or this is the orig. Nature Physics link with clickable citations:

🙏 We would like to thank our generous Patreon supporters who make Two Minute Papers possible:
Benji Rabhan, B Shang, Christian Ahlin, Gordon Child, John Le, Juan Benet, Kyle Davis, Loyal Alchemist, Lukas Biewald, Michael Tedder, Owen Skarpness, Richard Sundvall, Steef, Taras Bobrovytsky, Thomas Krcmar, Tybie Fitzhugh, Ueli GallizziIf you wish to appear here or pick up other perks, click here:

My research:
X/Twitter:
Thumbnail design: Felícia Zsolnai-Fehér –

#openai

Joe Lilli
 

  • @pandoraeeris7860 says:

    “It’s not every week you see an incredible AI breakthrough like this.”

    Correct. It’s every day (almost).

    XLR8!

  • @i3looi2 says:

    Everyone laughed when Tom Cruise was fidgeting his fingers in mid-air and saying “Enhance, enhance !” 😳

  • @Speak_Out_and_Remove_All_Doubt says:

    OpenAI needs to really sort out it’s naming schemes!

  • @torarinvik4920 says:

    I believe that when these AIs can start to create their own tools, even composing old tools to make new tools it could give incredible results. Perhaps using something program synthesis, or even training new neural nets that it can use for special tasks. Continuous learning is still too expensive to do with most current techniques, so would be cool to see some improvement there.

  • @lovegupta9262 says:

    Such an awesome video.

  • @louisrust says:

    you know it’s some serious stuff when the “dear fellow scholars” line is dropped 5 minutes into the video

  • @GIDAEONGULLY says:

    We can no longer mock “Computer, Enhance!”… RIP

  • @jean-francoiskener6036 says:

    honestly, gpt 1, 2, 3, 4 even 4.5, ok, i get it. But now you start with o3 and 4o i don’t have a clue what’s the last one. What’s more important, the “o” on the left, the “o” on the right, or the number? Who knows

    • @blackmartini7684 says:

      ​@@jean-francoiskener6036 o on the left. The 4o was 4 trained with images. o1 is a completely different model trained on reasoning itself. So o3 is the best and latest

  • @vaendryl says:

    we are so close to the level of “is it superhuman or just hallucinating?” and just have no way of knowing which it is.

  • @cfwin1776 says:

    Correction: it’s closeAI.

  • @AlvinCwk says:

    3 years ago 2minutepaper’s videos were about ai that can turn water into lava in a picture, now we are talking about ai that can do researches on its own…

  • @dasiudoi says:

    i just hope i live long enough for ai to extend my life.

  • @AnthonyPicciano says:

    All those movies that showed unrealistic tech…now: “Computer, enhance image” is real.

  • @lutzilutz9123 says:

    OpenAI marketing are so skilled, they must be laughing seeing science channel call them “pushing humanity forward”. Not there for the money at all. Wink wink

  • @davidtoomey1400 says:

    Thank you for your focus on tempered optimism and quality information. I purposely wait until you cover the topic. The fast news never compares the high quality content that requires time to produce. It’s worth the wait.

  • @TK621why says:

    Its looking up how to answer some of the questions on the “last exam” though, so not really understanding or having the reasoning skills necessarily

  • @woodland8281 says:

    Well done man. This continues to be great work.

  • @kipchickensout says:

    i gave it a pic of an engine and it correctly identified every part on it, as well as which engine it is and roughly of what car of what year range. I also gave it two street view images. on the first it got the correct town, likely because part of the street name was still visible. on the second one it was quite tricky but it correctly identified some smaller village in the very north of germany, likely in SH

    What a time to be alive!

  • @clivah1499 says:

    There will be two people with this research: People who turn into zealous crusaders to slay every machine, and people who will shed a tear and a smile at seeing this child of ours grow up.

  • @amarug says:

    This is all mindblowing and all, but it’s also heading with lightspeed towards the ultimate dystopia of total blandness. Imagine an AI can produce in real time an endless stream of the most amazing TV series tailored specifically to YOU. It will be better and more engaging than anything humans ever produced. Better music, better visuals, better acting, better everything and touch everything that YOU think is amazing. Extrapolate this thought to other fields etc and aside from things involving actual violence and deliberate harm to humans or animals, there are few things more nightmarish that I can think of. I think the fine line between “what a time to be alive” and “there is essentially no point in anything anymore” is getting thinner and thinner. I really am considering just buying a farm somewhere in the mountains here in Europe and live my life away from all of this. It will be a simple life but at least it’s going to be real… idk. I really worry.

  • >