When I asked something about a domain where someone is knowledgeable, I think that that person can ''reframe'' my question into a more interesting one by means of the very response she gave (she takes my question as a proxy for a better yet relate question). I think a chatbot cannot do that.

Someone more intelligent than me should think about it. There's something in the emergent culture of LLM usage that accelerates some trends that emerge with the internet. I don't know how to call it, the extent of it, but I can give an example: there's no more grandmother's secret recipe.

Yes. In my house we read the bible every morning. www.amazon.com/Self-Constit...

One funny thing for a foreign person about the political discourse in the usa is how much it is shape by bloggers.

Type 1: People who like LLM very much (are aware about they risk). They go bananas if you say LLM are stochastic parrots.

What's the provisional take on the First Proofs results? Are they more impressive than the IMO gold?

Altman did some effort to look like Oppenhaimer of our time (and for some moment he did get it), but now he is known as Sloppenhaimer. This is funny.

LLMs' discourse is kinda boring right now, or is it only me? They have undisputed and impressive applications in math, science and code, but I am bored as hell with AI in every news outlet. Even smart people are starting to have fewer & fewer cool things to say about them.

I like Bob Brandom very much; I like very much that he make avaible all his lectures; but I don't understanding why his lectures are so long. Is it some tradition in Pitt? I get the the conversational tone has clearly some point in it; but is difficulty to be in line with the reasoning display.

Unbelievable, I was listening on my Spotify John Cage 4'33 and it was an AI version.

'ChatGPT' is a nickname, and nicknames are for friends, and Chat Generative Pre-trained Transformer is no friend of mine.

We should only use words from the Bible, i.e. Tractatus Logico-Philosophicus.

Theodor Ludwig Wiesengrund-Enfeite.

Again, this precise use case is why these systems exist-to proliferate people’s racist (transphobic, misogynistic…) imaginary at scale. This should not be seen as a “misuse” but rather the product being used exactly as intended.

LAPD fired rubber bullets at Australian journalist @laurentomasi.bsky.social

Who's behind our new microfiche scanning livestream? None other than Sophia Tung @sof.is, the creative mind behind the viral robotaxi depot livestream! 🚕 Check out how Sophia pulled it all together, and why the work is so important at this moment in time ⬇️

this week's newsletter is about the lessons of our departed friend white genocide grok maxread.substack.com/p/regarding-...

Poisoning the air of a Black community for a fucking chatbot.

NEW: The music distribution platform SoundCloud updated its terms of service to allow it to use artists' uploaded music to train AI. The change quietly went into effect last year. It looks like a lot of SoundCloud users were caught off guard. futurism.com/soundcloud-a...

The newest and most powerful A.I. technologies — so-called reasoning systems from companies like OpenAI, Google and the Chinese start-up DeepSeek — are generating more errors, not fewer. As their math skills have notably improved, their handle on facts has gotten shakier.

I recently gave a public talk that mentioned this, and many people were genuinely surprised that the "free services" they used like Facebook, X, Google harvested their data to train AI models. The general public needs to be educated about this.

LLM in a nutshell.

And there it is. Putin’s lifelong dream. After years of relentless Russian propaganda, now fully embraced by the American president and the world’s richest oligarch, millions of Americans are convinced that this is a good plan because Europe is an enemy and Putin is our friend.

Wikipedia is now recognized as a digital public good by the UN-endorsed Digital Public Goods Alliance. This recognition highlights Wikipedia's role in advancing free and trusted knowledge, powered by nearly 260,000 volunteers in 300+ languages. Read more ➡️ wikimediafoundation.org/news/2025/02...