Acte I : l’empereur essaye ses nouveaux habits.

Nearly two years ago, OpenAI said that artificial general intelligence, or AGI — the thing the company was created to build — could “elevate humanity” and “give everyone incredible new capabilities.”

Now, OpenAI CEO Sam Altman is trying to lower expectations.

“My guess is we will hit AGI sooner than most people in the world think and it will matter much less,” he said during an interview with Andrew Ross Sorkin at The New York Times DealBook Summit on Wednesday. “And a lot of the safety concerns that we and others expressed actually don’t come at the AGI moment. AGI can get built, the world mostly goes on in mostly the same way, things grow faster, but then there is a long continuation from what we call AGI to what we call superintelligence.”

— Alex Heath, « Sam Altman lowers the bar for AGI », The Verge, 4 décembre 2024.

Interlude : le stylo est plus fort que la machine pensante.

Nilay’s main love language is trolling, and so he sent me this video of Sam Altman talking about note-taking, because he knew it would annoy me. […] I do find the video revealing. This is a man who has not carefully considered his tools and expects someone else to pick up after him. That does explain a lot about OpenAI, doesn’t it?

— Elizabeth Lopatto, « I have some notes on Sam Altman’s note-taking advice », The Verge, 25 novembre 2024.

L’article de Lopatto est hilarant sur la forme (« I am not going to get into the fountain pens question because it, like the paper discussion, is for sickos. »), mais inexact sur le fond. Même s’il gagnerait à remplacer son Uni-Ball Micro par un Pilot V5 et son carnet générique par un bloc Rhodia à spirales, Altman a choisi les outils les plus adaptés à ses besoins. Ce qui explique tout ce qu’il faut savoir sur OpenAI, c’est plutôt le comportement sociopathique qui consiste à jeter ses notes par terre plutôt qu’à la poubelle, parce que le monde entier devrait courber l’échine pour mériter un aperçu de l’infinie sagesse de sieur Altman.

Acte II : c’est dur d’être un AInsecte.

James and Rose, the bizarre AI bots who were recently installed as news broadcasters at local Hawaii paper The Garden Island, have been terminated. […] James, a middle-aged Asian man, and Rose, a younger redhead, were never able to figure out how to present the news in a manner that wasn’t deeply off-putting for viewers. Their program, which ran twice a week on Youtube, Facebook, and Instagram, covered topics as varied as a fall pumpkin giveaway and a vigil for a labor massacre—all in the same distant, matter-of-fact tone of beings incapable of comprehending human emotions.

— Guthrie Scrimgeour, « The AI Reporter That Took My Old Job Just Got Fired », Wired, 21 novembre 2024.

Acte III : faites l’AImour, pas la guAIrre.

OpenAI, the AI model maker that used to describe its mission as saving the world, is partnering with Anduril, a military contractor, the two companies announced Wednesday.

As part of the partnership, OpenAI will integrate its software into Anduril’s counterdrone systems, which detect and take down drones. It’s OpenAI’s first partnership with a defense contractor — and a significant reversal of its earlier stance towards the military.

Gaby Del Valle, « OpenAI is partnering with defense tech company Anduril », The Verge, 4 décembre 2024.

Épilogue : et puis il y a Siri.

Go figure — just one day after writing about how Apple’s ambiguous descriptions of supposedly clever features has the potential to rob trust, my phone has become haunted.

I saw a suggestion from Siri that I turn on Do Not Disturb until the end of an event in my calendar — a reservation at a restaurant from 8:30 until 10:00 this morning. No such matching event was in Fantastical. It was, however, shown in the Calendar app as a Siri Suggestion.

— Nick Heer, «  Siri Invented a Calendar Event and Then Hallucinated a Helpful Suggestion », Pixel Envy, 5 décembre 2024.

Tout va pour le mieux dans le meilleur des mondes.