I can literally listen to this guy on repeat.

This is incredible, I have so many questions.

Extraordinary claims that confirm your biases or trigger outrage are the easiest to believe, but that's exactly when skepticism matters most.

amazing how the entire machine can pivot within 24 hours. no need to slowly upsell the argument; just issue new marching orders

Bay leaves have to be a long-running experiment to see how long people will follow instructions without questioning them.

My 14-yo-son has recreated the climax of Back to the Future in gingerbread and I thought you should know.

Disney flavoured AI slop coming soon! openai.com/index/disney...

Just a reminder that Elon Musk is genocidal piece of shit. www.impactcounter.com/dashboard?vi...

Kurt Vonnegut’s definition of what makes something “art” is one of my favorites I’ve seen - from a letter to his brother in 1995.

Currently reading Chaos by James Gleick and loving it. @around.com

Just found out that Peter Thiele may be behind the push for the US to colonize Greenland, all in order to build a techno-utopian city called Praxis. Sounds like a bad conspiracy theory except that it's real. en.wikipedia.org/wiki/Praxis_...

Vous assistez, quasiment en direct à la formation d'un cratère sur la Lune, avec un impact filmé ce matin !

A 100x this

Was wondering about the impact of USAID cuts since I have seen much in the news, and found this. These people are monsters. www.impactcounter.com/dashboard?vi...

www.youtube.com/shorts/Fk4t_... #databs

15 years in the making, we confirmed that mitochondria - the powerhouse of the cell - have an unusual localization in patients who experience psychosis (including schizophrenia and bipolar disorders). You’ll never guess what kind of patient cells we used to make this discovery… 🧵

Why is linkedin just so very very bad.

Holy crap this belongs in a museum.

Always good to remember.

Found a scammer on FB creating ads, linking to a site that impersonates a local tabloid. Quite an elaborate scam, I guess LLM's make this easy. www.facebook.com/ads/library/...

New unit of measurement just dropped.

Seen this comparison several times in different forms. Comparing a specific grant which is an increase on TOP of total EU science funding, to US total funding. Sure the EU could be investing more but let's compare apples to apples. Also, is the EU expected to just instantly double it's budget? How?

Vector Sets are finally part of Redis! The full story here in my blog: antirez.com/news/149

"Life isn't about finding yourself, it's about creating yourself" - Bob Dylan

#timeseries #clustering #fails arxiv.org/abs/2503.14393 a great paper

Forget the 60k rows part. His "hard drive overheated" ?!

I know this is extremely vulgar idealism, but I do think we should try to aggressively reassert the norm that lying is bad. Not in a "Four Pinnochios" fact-checking kind of way, but in a way that frames this kind of lying as moral degeneracy that is fundamentally corrosive to the human spirit.

A friend included this anti-acknowledgement section on her PhD thesis. She also added the proposition: “Systematic bullying and undermining of girls and women in STEM starts early on and is the reason why they do not stay in science and related fields.” Absurd we still need to go through this 🧪👩‍🔬

Why does almost every electric car look like a tesla inside. Give me a normal car with knobs and cranks instead of a flats creen.

Now they're policing the speech of foreign universities. The party of free speech my ass. yle.fi/a/74-20147027

It's tempting to short Tesla given the current situation, but remember that it's a meme stock fueled by greed. Laws of logic don't apply here.

I just mistyped hyperlink as hyperkink🤔

If this goes ahead it will become a future case study on how not to use AI, both from a functional and ethical standpoint.

This may possibly be the "greatest harm" being carried out in modern history. Delaying an effective malaria vaccine by years will result in the unnecessary death of millions of children.

Dude-fluencers are selling "male victimhood" idealogy to privileged men, usually blaming women instead of the larger hyper captilistic system for their stress/inadequacy. A brilliant thread, research backed with a succinct framing of the issue we see all around us.

Regarding elon's salute shitfest. What I see missing from reporting and commentary is how he was obviously tweaked out of his fucking mind that day.

DeepSeek-R1! ⚡ Performance on par with OpenAI-o1 📖 Fully open-weight model & technical report 🏆 MIT licensed: Distill & commercialize freely! 🌐 Website & API are live now! Demo: chat.deepseek.com Models: huggingface.co/deepseek-ai

Lol what documentation? The X recommendation algorithm by now is probably a hairball of if-statements, each based on elon's random late night texts to devs.

Seriously, Fuck this guy.

Elon really does set the tone for all others to follow. Gut 80% of staff at Twitter and every tech company follows, at a smaller scale. Remove moderation under the guise of free speech and now Zuck follows. It doesn't matter how crappy the results of said changes were, they still follow him.

It's almost as if these prices were set by someone incredibly out of touch with what the plebs pay for things. Like the richest man in the world.

Something I don't understand is: why can't LLMs write novel-length fiction yet? They've got the context length for it. And new models seem capable of the multi-hop reasoning required for plot. So why hasn't anyone demoed a model that can write long interesting stories? I do have a theory ... +

Damn maybe Google an acronym before you choose it. Not even a deep search would have been needed.

It does piss me off that people are going to blame ML for killing the informativeness of the internet when twitter was bought by a billionaire, facebook led a pivot to video that seems to have decimated online journalism, and print publications are widely already killed, presumably by the internet.

I see posts talking about the use AI (I assume they mean llm's) as "deskilling". It's true to some extent but the same could be said for calculators. I suck at long division but that's free'd part of my mind to work on other stuff. #strongopinionlooselyheld

Meta has proposed a new type of Large Concept Model (LCM), which differs from traditional Large Language Models (LLMs) in two main aspects:

The law of accelerating goalposts.

Distrusting all media, including legacy outlets, often leaves us with far less reliable alternatives: grifter YouTubers, toxic Facebook groups, and viral WhatsApp messages.

o3 is impressive, let's see how long it takes to get costs down. I'm not spending 1K per pop to debug stack traces.