8 seconds vs 2 minutes 🤯

I put top 5 #LLMs to the test — #OpenAI o1, Google #Gemini, #Anthropic Claude, Meta #Llama, and #DeepSeek AI R1 — with a math problem:

𝐎𝐧𝐥𝐲 2 𝐠𝐨𝐭 𝐢𝐭 𝐫𝐢𝐠𝐡𝐭.

But the real insight wasn’t accuracy—it was speed! #AI Inference Time matters just as much as Inference cost!

Hot off the presses! 🔥📰

The new Trending News feed takes the 📰 News feed and sorts it by popularity. This is a work-in-progress I've been developing with @devingaffney.com, more info soon!

bsky.app/profile/did:...

The #Gladiator 2 #musical we need, but don't deserve!

So much of today’s #AI and #LLM wave is due to #GPUs. But it's not always obvious how much GPUs really outperform CPUs.
This #Mythbusters video from way back in 2009 (yep, 15 years ago!) is such a practical way to explain extreme parallel computing 🧮.

Finally made it to BlueSky! 👋