Besides model quality, it has been incredible to see the advances in infrastructure that happened in the last year (efficiency, speed, easier to use & ship). I am sure things are still far from perfect, so please give (constructive) feedback to @clmt and @OfficialLoganK
@oriolvinyalsml
-
Model Self-Improvement: Emergent Behavior in AI Systems
By
–
Check out this example which showcases one of the most exciting research directions: self-improvement.
— Oriol Vinyals (@OriolVinyalsML) 12 décembre 2024
In it, you see this behavior emerging (!) when the model realizes (with a "Oops!") that it did a mistake, and fixes it to create the cute image.
Wild times. https://t.co/bSwfUa7QCD pic.twitter.com/TasWGsgn5sCheck out this example which showcases one of the most exciting research directions: self-improvement. In it, you see this behavior emerging (!) when the model realizes (with a "Oops!") that it did a mistake, and fixes it to create the cute image. Wild times.
-
Google Gemini 2.0 Flash Surpasses 1.5 Pro with Native Image Generation
By
–
Gemini 2.0 Flash ⚡️ has arrived!
— Oriol Vinyals (@OriolVinyalsML) 11 décembre 2024
2.0 Flash > 1.5 Pro (again!) 📈
Interacts with a browser 🤖
Native image generation 🖼️
and much more!
Try it out https://t.co/IMN3bqKgJS
As a preview of what is possible, wishing you all a Drastic Holiday powered by 2.0! pic.twitter.com/iHBB4lwXTkGemini 2.0 Flash has arrived! 2.0 Flash > 1.5 Pro (again!)
Interacts with a browser
Native image generation and much more! Try it out
http://
aistudio.google.com/prompts/new_ch
at
… As a preview of what is possible, wishing you all a Drastic Holiday powered by 2.0! -
Pareto Analysis Visualization with Significant Data Points
By
–
I drew the Pareto for you. Lots of red : )
-
Gemini Anniversary: Google Competitive AI Progress Update
By
–
A year ago we launched Gemini Back then I shared with the team that it would be great to be in a position to beat ourselves. Just a year later this is starting to happen. But we have formidable competitors, so I expect this to age poorly . GL HF! https://
aistudio.google.com/app/prompts/ne
w_chat?model=gemini-exp-1206
… -
Scaling Simple Recipes: Researchers’ Perspective on AI Model Growth
By
–
Scaling up a simple / boring recipe is not what many researchers wanted to hear at the time.
-
Sharing Prize with GANs Generative Adversarial Networks
By
–
Yup great to be sharing the prize w/ GANs!
-
NeurIPS Honor for Transformers Research and Large Language Models
By
–
Such a great honor, thanks a lot @NeurIPSConf and congrats to my esteemed co-authors @ilyasut & @quocleix
! The 2014 talk also stood the test of time IMO. Here is a slide from it (powerful models of today == large transformers). Believe it or not this talk was controversial at the -
AMA on Large Language Models While Gaming
By
–
A dream of mine: do an AMA on LLMs whilst streaming WoW gameplay
-
AI Evolution: From SIFT to LLMs and Beyond
By
–
Raptor 1: Traditional ML: SIFT + SPM + SVM
Raptor 2: Deep Learning: min_θ loss(NeuralNet(data, θ))
Raptor 3: LLMs: "Please solve this coding puzzle"
Raptor 4: ???