(Also very much a missed opportunity for Google to rename hallucinations here, which is a terrible name for it.)
@mattlynley
-
Data Quality Over Quantity: Gemini’s Intentional Uncertainty Handling
By
–
Data quality > data quantity, as we've been talking about since september :). Fun part here, it includes examples where Gemini intentionally throws up its hands (like another model out there). You'd be surprised how useful people think this is. https://
supervised.news/p/ai-in-septem
ber-reasoning-engines
… -
Google’s JAX AI Framework Undergoes First Major Test
By
–
Google's new home-grown AI infra framework JAX getting its first mega-test here.
-
Google adopts Meta’s approach training smaller models massive datasets
By
–
And Google now running with Meta's approach of training smaller models with massive data sets. If you're curious about size, the open source RedPajamas open data set is 30T tokens. We can prob safely assume Google has access to much more.
-
Nano Gemini Model Quantized at 4-bit Resolution
By
–
As expected Nano is a distilled Gemini model quantized (at 4-bit, interestingly)
-
MTEB Evaluation for Bedrock and Vertex Embeddings
By
–
Has anyone done an MTEB eval for bedrock or vertex embeddings?
-
Non-obvious AI developments beyond major corporate shake-up
By
–
For Friday’s issue, a recap of some of the non-obvious stuff that happened in AI outside of that one guy getting kicked out of the office for a bit
-
Cloud Compute Consumption Shift Driven by AI Growth
By
–
The other big shift we're seeing is a change in consumption in cloud compute. Snowflake's earnings signaled that there would be a rebound in consumption based pricing, and you hear about it more these days among startups. AI isn't the whole story, but it seems like a big part.
-
Embeddings and RAG Models Driving Major AI Improvements
By
–
But one of the hottest areas by far is going to be embeddings. RAG is becoming a fixture in AI to improve the quality of model responses, and improvements in embedding models is going to be a big part of that. And after months of crickets, we got three big launches in 45 days.
-
OpenAI’s Major Event Sparks Wave of AI Challenger Startups
By
–
First, the obvious, that thing at OpenAI. It was a big deal and probably the most consequential story in tech in the last decade. But it also opened the door for challenger startups—29 of them that I'm tracking closely, to be exact.