2/We've seen 7 major models from top labs in the last 3mo: May:
– GPT 4o
– Gemini 1.5 Pro June:
– Claude 3.5 Sonnet July:
– Llama 3.1
– Mistral Large 2
– GPT-4o Mini August:
– Gemini 1.5 0801 Each of these models has been incredibly competitive—each world-class in some way.
RESEARCH
-
Seven Major AI Models Released in Three Months
By
–
-
Gemini 1.5 Pro Tops LMSYS: Google’s Compute Edge Advantage
By
–
1/Gemini 1.5 Pro 0801 is the new best model (tops LMSYS, SEAL evals incoming) Key considerations
1—OpenAI, Google, Anthropic, & Meta all right ON the frontier
2—Google has a long-term compute edge w/TPUs
3—Data & post-training becoming key competitive drivers in performance -
New KV Cache Framework Integrates Latest Research Papers
By
–
For those of you that can't keep up with all the new stuff from @answerdotai
, I've got bad news–we have more! @GriffinAdams92
's first project is a KV cache framework that brings the latest research papers together. (And rather than write a blog post, he wrote an epic tome.) -
OpenDevin: Open Source Generalist LLM Agents Framework
By
–
Open source generalist LLM Agents! Excited to have author @xingyaow_ on alphaXiv this week to answer questions about his recent paper OpenDevin.
-
Google DeepMind’s Gemini 1.5 Pro Tops LLM Leaderboard
By
–
@GoogleDeepMind takes the lead with Gemini 1.5 Pro for the first time on the @lmsysorg leaderboard! We're excited to try this model out!
-
Pet Cloning in Spain: Genetic Replication Cannot Replicate Identity
By
–
La clonación de mascotas ha llegado a España y por el astronómico precio de 55.000€ puedes obtener una réplica genéticamente exacta de tu amigo peludo.
— Juan Merodio (@juanmerodio) 1 août 2024
Pero por mucho que por fuera tu perro sea igual, el animal nunca será el mismo. Hay muchos factores que influyen en el… pic.twitter.com/LfEBECDATTLa clonación de mascotas ha llegado a España y por el astronómico precio de 55.000€ puedes obtener una réplica genéticamente exacta de tu amigo peludo. Pero por mucho que por fuera tu perro sea igual, el animal nunca será el mismo. Hay muchos factores que influyen en el
-
Beyond Accuracy: Rethinking AI Model Evaluation and Assessment
By
–
If you're thinking about this as an "accuracy" problem, you've already lost the plot … as well illustrated by this quote from one of the dissenting voices hidden in the middle of the article:
-
Medical Science Shouldn’t Platform Automating End-of-Life Care
By
–
This is the same horrific idea that I wrote about in my MAIHT3k newsletter post on Monday, but @techreview goes for the "ooh shiny!" angle instead: https://
buttondown.email/maiht3k/archiv
e/medical-science-shouldnt-platform-automating-end/
… >> -
MIT Tech Review’s Mixed AI Coverage: Ethics vs. Tech Boosterism
By
–
The reporting in @techreview is such a mixed back. Some good critical work, some "look shiny!" tech reporting that at best quotes a few "skeptics" towards the bottom before returning to boosterism. This one is an example of that — a short thread: https://
technologyreview.com/2024/08/01/109
5551/end-of-life-decisions-ai-help/
… >> -
Gemini 1.5 Pro Reaches #1 on LMSys Leaderboard at 1300 ELO
By
–
Our experimental version of Gemini 1.5 Pro in AI Studio debuted at #1 on the LMSys leaderboard. 1300 ELO, nice round number With the Gemini team, we are focused on both improving our models by iterating and finessing our recipe (deep learning FTW!), and also on massive bets