Apple lance une application Siri indépendante et dote l'iPhone d'une mémoire permanente

Eduardo Bolsonaro🇧🇷 @BolsonaroSP

🇧🇷🇺🇸 It’s official: Lula’s Federal Police have now placed the United States under investigation. The narrative is setting the stage so that, in the event of a @FlavioBolsonaro victory, the election can be annulled under the pretext of “foreign interference.” And this is happening even without USAID money pouring into Brazil, without the CIA director and other senior U.S. officials coming to Brazil to pressure Brazilian authorities. In other words, without anything resembling what already happened with the Biden Administration’s support for Lula in 2022.

Pirat_Nation 🔴 @Pirat_Nation

A new tool called SCSKiller is trying to get rid of one of the most annoying problems in PC gaming: shader compilation stutter. The tool pre-compiles shaders before you play, so games don’t have to stop and do the work during gameplay. Wccftech tested it with Final Fantasy VII Rebirth and the results were pretty impressive. After using the tool, the shader-related stutters they encountered were basically gone. The project also reports some big improvements in other games: Final Fantasy VII Rebirth: 19 → 0 stutters Hogwarts Legacy: 1,014 → 87 Star Wars Jedi: Survivor: 9,038 → 11 SCSKiller currently works with NVIDIA and AMD GPUs on DirectX 12, while Intel GPUs aren’t supported yet. https://wccftech.com/new-scskiller-tool-targets-pc-shader-compilation-stutter-our-final-fantasy-vii-rebirth-test-shows-impressive-results/

Theo - t3.gg @theo

July 2026: Anthropic has the best code models. Gap isn’t very big though. They’re slow, expensive, and the “claudeisms” are at an all time high. And your $200 sub is so limited that you can kill it in a day. Preferring OpenAI models made a lot of sense. You lose a slight edge on capability, but they’re faster, cheaper and more pleasant to use. Your $200 sub feels unlimited, and you get resets from Tibo every 3 days on average September 2026: Anthropic has the best code models. The gap is way bigger now. Opus 5.5 is fast, surprisingly cheap, and the prose is pleasant to read. Your $200 sub suddenly feels nearly limitless. Preferring OpenAI models for code makes almost no sense. You lose a lot of capability and reliability. They’re slow as hell without fast mode, which makes the price far too high. You can kill your $200 plan in a few hours, and the $500 plan goes even faster with UltraFast. I know OpenAI can come back from here, but it’s looking rough right now.

Fox News @FoxNews

JUST IN: @karolineleavitt is joining Fox News Channel as an on-air contributor after leaving the White House briefing room. The former White House press secretary will provide commentary across the Fox News lineup beginning Nov. 1. "Karoline’s firsthand experience in the White House and on the presidential campaign trail gives her a distinct understanding of today’s political landscape," FOX News Media CEO Suzanne Scott said, adding that Leavitt will bring "authentic and sharp insight" across the network's platforms. Leavitt says she's ready for the next chapter: "I look forward to bringing my unique perspective from my experience as White House press secretary, along with the same energy and candor that defined my tenure, to Fox’s millions of viewers across their many platforms." Leavitt stepped down from the White House in August after making history as the youngest press secretary, saying she wanted to devote more time to her two young children. Read more about Leavitt's move to Fox News and when viewers will start seeing her on air: https://www.foxnews.com/media/karoline-leavitt-joins-fox-news-contributor-leaving-white-house-press-secretary-role

Social Capital @socialcapital

Microsoft CEO Satya Nadella wrote that a company renting an AI model “essentially pays for intelligence twice”: Once in money, and again in the proprietary knowledge it has to reveal to make the model useful. This mechanism is called the learning loop: every prompt and correction captures some of the company's know-how. That knowledge is distilled and, with a closed model, the provider keeps that information. If your usage data is worth more than the token cost to serve, providers might end up paying you to use their models. The precedent exists: OpenAI created an opt-in program that gives accounts up to 1M free tokens a day on certain models in exchange for training rights on that traffic. Chips can be rented and algorithms can be leaked or distilled, but the prompts and corrections flowing through a model compound privately for whoever holds them.

Commentary Donald J. Trump Truth Social Posts On X @TrumpTruthOnX

125 Million People voted today in the first round of India’s Presidential Election. The results came out shortly after the vote closed, on the SAME DAY! In Detroit, Philadelphia, California, and numerous other U.S. Cities and States, the results, with much smaller numbers, will often take weeks to Rig, I mean, calculate. Voting in America is CORRUPT!!! President DJT ( TS: Oct 4 2026, 11:32 PM ET )​​​​​​​​‍‌​​‌​​‌​‍​‌​​​​‍‌‌‍​‌‍‌‍‌‌‌‍​‌‌‌​‍

Pirat_Nation 🔴 @Pirat_Nation

A modder has managed to get Nvidia Pascal GPUs running on 32-bit Windows XP with unofficial drivers. The modified drivers add support for cards like the GTX 1060, 1070, 1080, 1080 Ti and Titan Xp, while also improving compatibility with modern monitors through better DisplayPort and HDMI support. Pascal never officially supported Windows XP, so this is entirely a community-made solution aimed mainly at retro PC enthusiasts.

*Walter Bloomberg @DeItaone

$SPCX - MORGAN STANLEY: SPACEX “CHEAP AND GETTING CHEAPER” Morgan Stanley reiterates Overweight on SpaceX with a $300 price target, arguing the stock looks cheap once its growth profile is considered. Analyst Adam Jonas says SpaceX trades roughly 40% below mega-cap AI peers on a growth-adjusted basis. Potential catalysts include upcoming Starship flights, Q3 earnings, Grok releases and additional neocloud deals, with a successful Starship catch potentially the biggest catalyst since the IPO.

Tyler Glaiel @TylerGlaiel

Is there any actual writeup about how those AI generated game mashups are actually working? There are bits of this that aren't quite passing the sniff test for me

Chubby♨️ @kimmonismus

Looks like everything is being rebranded as "Super Ingelligence" instead of "Artificial Intelligence". Even SpaceXAI will be named SpaceXSI