@elonmusk@growing_daniel why you still trying to make opeanai doing dumb? please agree that they are good and grok cannot achieve that performance for now.
A simple yet powerful example of the new Gemini 2.0 Flash's native multimodal input + output. Precise conversational editing & reasoning!
Next step, Chess!
Within 24 hours, @GoogleDeepMind has released a Quantum Chip, that "performed a standard benchmark computation in five minutes that would take one of todayβs fastest supercomputers 10 septillion (that is, 10^25) years β a number that vastly exceeds the age of the Universe." and @openai has released Sora, which gives anyone the ability to create videos that look real but are fictional in superior quality.
Within 7 days we got:
- with GenCast from Google DeepMind a weather model that generates weather forecasts in unprecedented quality (15 days in advance)
- with Genie 2, a world model that can train agents that even observes physical laws and stores the environment within
- with LLama 3.3 70b an open source model that is on par with a model that is almost 6x as big and was released only 5 months ago (Llama 3.1 405b)
- OpenAI o1 Pro, a reasoning model that has been described as outstanding by numerous scientific experts
I don't even want to list all the things we've received in a year because it's too much, but just remember AlphaFold 3, SearchGPT, CursorAI, Sonnet 3.5 and so much more.
The acceleration continues to increase, it's hardly possible to keep up with everything.
This is the best time in human history to be alive. We are witnessing the approaching singularity.
I love it.
Time to finally read this one. My lead engineer recommended it to me. Being web devs, we often overlook what our memory does. This serves as a good introduction to how memory works.