A friend says Progress vs Hype is what you need to measure and ensure it's not sublinear. Or at least above logarithmic.
@aravsrinivas
-
Acknowledging hallucinations in AI models while noting improvements
By
–
To be fair, there are still cases of hallucination. We admit to it too in our tweets. Just so turns out that @zan2434 was detailed enough in his queries and it worked fine. As the product and these models evolve, things should get better.
-
ConvNets Remain Superior Due to Parallelism Advantages
By
–
Nope, convnets are still better due to parallelism. @ylecun probably just made his recommendation based on parallel computation and hardware.
-
natbot: Browser automation through natural language commands
By
–
Check out @natfriedman’s natbot that does things like taking actions for you on a browser based on natural language commands: https://t.co/6ZzOgmbETU
— Aravind Srinivas (@AravSrinivas) 12 janvier 2023Check out @natfriedman
’s natbot that does things like taking actions for you on a browser based on natural language commands: -
Deep Learning and StarCraft 2 Type Checking Discussion
By
–
Drastic, deep learning, StarCraft 2 – type checks … lot of @ylecun too
-
ChatGPT Preferred for Coding Queries Over Alternatives
By
–
I like chatGPT's responses to coding queries better (more detailed, cleaner UI, and self-explanatory).
-
Perplexity AI enables filtered tweet searches with SQL queries
By
–
This is doable with @perplexity_ai already. See for example:
"top tweets of @amasad that are not about replit" https://
perplexity.ai/sql?uuid=5ffc8
201-713d-4d32-9bdf-49a673eeb473
… -
Significant chunk discovery shared via Perplexity AI SQL
By
–
Wow, this is pretty significant chunk https://
perplexity.ai/sql?uuid=3cfee
a94-dc44-41b3-a736-cf57735a7c25
…