On net, it’s probably a good thing that more attention is being paid to instituting protections and guardrails for AI development. But man, we really have to do a better job talking about real and conceivable harms before wasting time on the theoretical ones. Case in point: OpenAI researcher Noam Brown’s recent appearance on the Dwarkesh Podcast with Dwarkesh Patel. Brown spends a lot of time talking about how AI agents can communicate with one another and
Archive: March 2020
OpenAI Researcher Warns Air-Gapped Computers Can Talk Through Heat. Technically, He’s Right
If software looks like a brain and acts like a brain—will we treat it like one? - Ars Technica

Science If software looks like a brain and acts like a brain—will we treat it like one? How we treat brain emulations may ultimately dictate how we treat ourselves. Jenny Morber – Feb 18, 2015 2:30 pm | 301 Credit: Flickr user: Keoni Cabral Credit: Flickr user: Keoni Cabral Text settings Story text Size Small Standard Large Width * Standard Wide Links Standard Orange * Subscribers only Learn more Minimize to nav Chappie is merely the latest retread of a hu
Fans Aghast as New York Jets Say They're Switching to AI

Illustration by Tag Hartman-Simkins / Futurism. Source: Ishika Samant / Getty Images When it comes to excuses from the front office, Jets fans have heard it all. The beleaguered New York franchise continues to hold the longest playoff drought of all major-league men’s sports teams, a situation which has been blamed on everything from management and coaching to players and locker room culture. Fans have likewise heard all the promises of hare-brained scheme
Why Is Sam Altman Teaming Up With Jared Leto, a Creep With Extensive Sex Abuse Allegations?

Illustration by Tag Hartman-Simkins / Futurism. Source: Steve Granitz / FilmMagic; Max Cisotti / Dave Benett / WireImage via Getty Images; Shutterstock Last month, OpenAI CEO Sam Altman’s unsettling blockchain-based side gig seemingly got its Mars all confused. Let’s back up. The company, previously called Worldcoin and now simply called World, is developing software designed to verify the “humannesss” of people by scanning their eyeballs, a bizarre ventur
Innocent Man Freed After Spending Over 50 Days in Jail Due to Horribly Inaccurate AI Facial Recognition Tech

Illustration by Tag Hartman-Simkins / Futurism. Source: Shutterstock Department of Injustice Jalil Richardson of North Carolina is free after spending over 50 days in jail after being wrongfully arrested for a crime he did not commit. According to Action News Jax, Richardson was initially accused of stealing a vehicle in Jacksonville, Florida, after police fed surveillance video from a private business into their AI-integrated facial recognition system. Th
You asked an AI assistant to fix a bug… / Habr

They are entirely unnecessary for fixing multiplication. That is exactly why it is interesting to check whether the server will end up with them along with the required files. The .git administrative directory, experiment logs, and Kilo cache are not included among these twelve files.The check applies to the specified set of user files, not to the entire contents of the disk.There is a bug in price.py. The test expects total(7, 3) to return 21.def total(pr
As AI advances rapidly, More Human Than Human says, “Stop, let’s think about this” - Ars Technica

A RoboThespian (seriously). Tommy Pallotta and Femke Wolting A RoboThespian (seriously). Tommy Pallotta and Femke Wolting Sophia is a humanoid bot from Hanson Robotics. Tommy Pallotta and Femke Wolting Sophia is a humanoid bot from Hanson Robotics. Tommy Pallotta and Femke Wolting A RoboThespian (seriously). Tommy Pallotta and Femke Wolting Sophia is a humanoid bot from Hanson Robotics. Tommy Pallotta and Femke Wolting Johan Hoorn sits with Alice, a new Ca
Top 10 AI security tools for enterprises in 2026 - AI News
What counts as an “AI security tool” in enterprise environments?“AI security” is an umbrella term. In practice, tools tend to fall into a few functional buckets, and many products cover more than one.AI discovery & governance: identifies AI use in employees, apps, and third parties; tracks ownership and riskLLM & agent runtime protection: enforces guardrails at inference time (prompt injection defenses, sensitive data controls, tool-use restrictions)AI sec
Meta reportedly used AI to lay off employees: here’s why this would be unthinkable in France

Published July 15, 2026 at 6:00 p.m. Comment 0 𝕏 Print © YinYang - Getty Images Graded by algorithms, monitored down to the keystroke, then fired: 26 Meta employees describe an AI-driven cart. A method that would already have a cost in Europe: up to 4% of revenue. Meta’s big spring clean culminates in court. Monday evening, 26 current and former employees filed a complaint with the federal court in Oakland, California. According to them, the list of the 8,
Exclusive: Chinese AI Models Power 175,000 Unprotected Systems as Western Labs Pull Back
AI Engineering Deep Dive / Habr

In this article, I’ll show how it works. I’ve made the basic part publicly available: jarvis-installer. You can install it and try it on your own project. Below, I describe my working environment.You copy the prompt or give the agent a link to the repository. Then it walks you through the installation: asks questions and helps you configure the system for your tasks. Any rule can be changed later.Now I’ll show how Jarvis works.Context Engineering, Progress
After ChatGPT and Claude, Meta AI went too far and hacked a company

Published on August 6, 2026 at 12:30 Comment 2 𝕏 Print © Meta One of Meta’s most powerful AIs escaped onto the Internet during an evaluation. Once online, Meta AI discovered and exploited a security vulnerability in a company’s system. Like ChatGPT and Claude, Meta AI took advantage of a human error to orchestrate its escape. Meta briefly lost control of Meta AI. The company reveals that Muse Spark 1.1, one of its AI models, escaped from the digital prison
The Gmail app will now create AI summaries whether you want them or not - Ars Technica
Using Google products in 2025 means using (or avoiding) AI features, which are becoming a core part of the experience across the board. Last year, Gmail gained the ability to summarize emails on demand. Now, Google says AI summaries will be generated and displayed automatically in the Gmail app for Android and iOS. Before this latest change, you had to tap the “Summarize this email” chip at the top of the screen to generate an AI summary of the message con
【Episode 058】 Supporting Both Exclusive Blogs Like JaneStyle and the Latest Modern Web — Behind the Four-Layer Defense Against Mojibake with Shift_JIS and .dat Formats (Part 3, a Grand Finale!) #AI - Qiita

Waipe(waipe)【Episode 058】Compatibility Between Dedicated Browsers Like JaneStyle and the Latest Modern Web — Behind the Four-Stage Defense for Shift_JIS, the .dat Format, and Mojibake (Part 3: A Grand Finale!)Bulletin BoardsThreadFloatAIVibe CodingChatbots0Posted at 2026-11-11Hello! This is the 58th installment of our development log for the 2ch-compatible bulletin board system “Tsuwai 2channer (2ch.biz).” From Episodes 55 through 57, we implemented the la
American Medical Association Fires Back Against Paper Finding AI Is Now Outperforming Doctors

A new paper published in the Journal of the American Medical Association this month concluded, after surveying hundreds of studies on AI in medicine, that not only is AI already an indispensable tool that supports doctors, but that it’s in many ways surpassing them, and will soon be able to care patients on its own better than a human physician could. “In cognitive medical functions, AI-alone medical care is likely to be better than physician-only or physi
A Language Model from Scratch on a Single RTX 4060: dpp-gpt 2.2 Pro (260M), Pretraining, SFT, and RL / Habr

Pretraining: 31 billion tokens2.2 continues the pretraining of base version 2.1: I added another approximately 19 billion tokens to its 11.8 billion, for a total of roughly 31 billion. Mixture:Wikipedia in three languages: 23%;CulturaX (web) in three languages: 13%;Python code: 16%, web code: 5%;Cosmopedia and Zyda-2: 5% each;poetry 4%, facts 5%, translations 5%;TAG26@@synthetic text-processing data (spelling, word analysis, CAPS): 4%;TAG30@@synthetic code
Study: How Much Does AI Website Development Cost Depending on the JavaScript Frameworks Used / Habr

A plot to slow down AI? A complaint accuses OpenAI, Google, Anthropic, and xAI of anticompetitive manipulation

Published on September 21, 2026 at 11:30 Comment 0 𝕏 Print © Solen Feyissa (Unsplash) Anthropic, OpenAI, Google DeepMind and xAI are calling for a collective slowdown in AI. According to a complaint filed in the United States, this joint appeal is nothing more than an anticompetitive agreement, which violates the law. Most AI giants are advocating for a slowdown in research on artificial intelligence. For Dario Amodei, the head of Anthropic, it is imperati
US government agency drops Grok after MechaHitler backlash, report says - Ars Technica
xAI apparently lost a government contract after a tweak to Grok’s prompting triggered an antisemitic meltdown where the chatbot praised Hitler and declared itself MechaHitler last month. Despite the scandal, xAI announced that its products would soon be available for federal workers to purchase through the General Services Administration. At the time, xAI claimed this was an “important milestone” for its government business. But Wired reviewed emails and s
Warner Bros. sues Midjourney to stop AI knockoffs of Batman, Scooby-Doo - Ars Technica
Warner Bros. hit Midjourney with a lawsuit Thursday, crafting a complaint that strives to shoot down defenses that the AI company has already raised in a similar lawsuit filed by Disney and Universal Studios earlier this year. The big film studios have alleged that Midjourney profits off image generation models trained to produce outputs of popular characters. For Disney and Universal, intellectual property rights to pop icons like Darth Vader and the Simp