Pain is the new weight loss
This took only six months for the FDA to approve. https://www.biopharmaboardroom.com/news/11/2108/fda-grants-priority-review-to-vertexs-suzetrigine-targeting-january-2025-for-action-date.html It blocks nerve signals for pain. It’s a pill. The implications are going to be wide and varied.
Pain is the new weight loss
This took only six months for the FDA to approve. https://www.biopharmaboardroom.com/news/11/2108/fda-grants-priority-review-to-vertexs-suzetrigine-targeting-january-2025-for-action-date.html It blocks nerve signals for pain. It’s a pill. The implications are going to be wide and varied.
DeepSeek or DeepFake
Distillation is basically “compressing” a model. Perhaps someone else’s model. And that’s a lot cheaper than making a model in the first place. Here’s my post from a couple days ago about why DeepSeek was cheap to make – because
DeepSeek or DeepFake
Distillation is basically “compressing” a model. Perhaps someone else’s model. And that’s a lot cheaper than making a model in the first place. Here’s my post from a couple days ago about why DeepSeek was cheap to make – because
Why was DeepSeek so cheap to make?
My summary: raw data not labeled data or human aligned, curated data collections given privilege (not boil the ocean), train against other-people’s-models, “mini-models” that are trained by the big model = cheaper to train, cheaper to inference. They refine/scaffold off
Why was DeepSeek so cheap to make?
My summary: raw data not labeled data or human aligned, curated data collections given privilege (not boil the ocean), train against other-people’s-models, “mini-models” that are trained by the big model = cheaper to train, cheaper to inference. They refine/scaffold off
Did GM even try to make flying cars?
Blade Runner, set in a Tokyo-like place, in 1982. GM did a bit of work in the 60s with NASA. But I guess they got distracted. Also – watch Blade Runner 2049. Many issues we will face soon about our
Did GM even try to make flying cars?
Blade Runner, set in a Tokyo-like place, in 1982. GM did a bit of work in the 60s with NASA. But I guess they got distracted. Also – watch Blade Runner 2049. Many issues we will face soon about our
New data center math: cheaper models, lazy inference, mini parameters = save $
Kluster.ai bands together data centers and unique deployments to provide more efficient compute. Try it for DeepSeek https://www.kluster.ai/#pricing
New data center math: cheaper models, lazy inference, mini parameters = save $
Kluster.ai bands together data centers and unique deployments to provide more efficient compute. Try it for DeepSeek https://www.kluster.ai/#pricing
At Davos last week, the only person to say “DeepSeek” was the Chinese professor from Tsinghua
But everyone else was well aware: Gary Marcus, Demis, Tegmark, OpenAI, Amadei, plus the reporters (WSJ, Bloomberg). Why is it news THIS week? The stuff was released Dec 26.
At Davos last week, the only person to say “DeepSeek” was the Chinese professor from Tsinghua
But everyone else was well aware: Gary Marcus, Demis, Tegmark, OpenAI, Amadei, plus the reporters (WSJ, Bloomberg). Why is it news THIS week? The stuff was released Dec 26.
DeepSeek, plus head to head
How big a deal is DeepSeek from China? Well it could invert the Nvidia Curve
DeepSeek, plus head to head
How big a deal is DeepSeek from China? Well it could invert the Nvidia Curve
You must be logged in to post a comment.