How many Smol simps are there ? retweet for free mint 👀

An amazing breakdown of the Wormhole hack.

Stay safe out there everyone. Watch out for the scam account impersonating adri.algo | Adri Belotti

Join Karen Cohen to learn more about #AlgoHUB's reach Bounty Hackathon on Tuesday, March 8. Build in Reach, launch on Algorand Technologies, win amazing prizes. meetup.com/meetup-group-C…

Mind blowing how technology compounds over time!

😉 Not part of the thread, but thought I mention to the #algofam that I'll be interviewing Silvio Micali Algorand Technologies on Day 2 11am AEDT. Make sure to get a virtual ticket and tune in. Later in the day Daniel Oon Algorand Foundation will join the "Web3 - The building blocks" panel 🚀

It's Time to program on the #blockchain- Reach Platform Hackathon Kick Off 12.30pm AEST Tuesday 29 March earn The engineer program like a boss Blockathon_DAO - 区块松DAO algorandfoundation Karen Cohen NFTs For Good #devs #NFTs meetup.com/en-AU/meetup-g… #Meetup via Meetup

The final zoom out broke my brain 🚼

Amazing work! Google maps street view in full VR seems closer than it ever has

I mean, I wouldn’t describe it as “uncontrollably”…

Unsure if real prompts or hallucination of the chat agent, but interesting!

the Hugging Face team is doing such amazing work!

Lool! One of the best Shit User Story yet.

I've put my lecture notes from my Fall 2022 Quantum Computation course online. math.mit.edu/~shor/435-LN/ I don't guarantee that there aren't any mistakes left in them, although I've tried to eradicate them all. If you find any, feel free to email me.

When OpenAI dropped the cost of o3, and in some cases it TPs seemed to improve, it's like because they moved them to google's TPU's right? Assuming the TPUs have a similar efficient to groq's LPUs (which make super fast serving of models relatively affordable), it makes sense.

I assume this is so they can save money/increase speed on inference using TPUs to run ChatGPT to expand free services and reduce costs. As the only other architecture similar and competitive is Groq, and there is no way they could scale that many chips afaik.