Jonathan Teutenberg

Visions of Nature | Natural History Museum Set 100 years into the future, our first mixed reality experience combines innovative technology and with expert science to imagine a world within reachwhere people and planet are thriving.

Augmented reality at the Natural History museum worked for a 10 and 13 www.nhm.ac.uk/visit/exhibi...

Anthropic's Subliminal Learning: Language Models Transmit Behavioral Traits via Hidden Signals in Data

When an LLM subliminally learns a pattern—one that may not be directly labeled, obvious, or even intentionally present—fine-tuning a student model on the outputs of that LLM enables the

I think they are still technically bao - probably "xiao long bao" - big in Shanghai.

This is the greatest creation I have ever seen from GW. Full marks.

"Firstly, can someone have the intern go pick up a roll of infinite tape..."

Also 99p on Kobo right now. Bought.

Does it not bother you that both the fool-er and the fool-ee are referred to as the fool? This whole day is ripe for confusion :/

Even if the total load at a charging station is the same, there can be a big improvement for users.

One charger being vacated 10x faster results in about half the average waiting time compared to 10 chargers in parallel. And much less if it's not crowded.

That checks out.

I recall seeing a photo from a nightclub with a man explaining proof of work by prime factorisation to a lady.

Hi, so I've spent the past almost-decade studying research uses of public social media data, like e.g. ML researchers using content from Twitter, Reddit, and Mastodon.

Anyway, buckle up this is about to be a VERY long thread with lots of thoughts and links to papers. 🧵

Now this place is cooking, maybe worth doing a big old energy thread.

Let’s talk about how electricity prices are set, why they’re so high in Britain, and how we could reduce them…
🧵

Just finished part 1 and as someone who studied AI through one of its winters and is now in the thick of it, I thoroughly endorse this podcast.

Plenty of (necessary) simplifications and omissions but nothing wrong on a pretty technical topic. Nice job.