Pew Research: - 64% of U.S. teens aged 13–17 are already using AI chatbots - 54% use them for schoolwork - 47% for fun and creativity What do you use AI for in your daily life? pewresearch.org/internet/202…

OpenAI strikes in the talent war: poached Ruoming Pang – former head of models at Apple who joined Meta only 7 months ago Meta lasted just seven months… brutal Real boost for Sam or just another hire? theinformation.com/briefings… #AI #OpenAI #TalentWar

Google drops Nano Banana 2 – new image generation model Claims Pro quality at Flash speed, strong consistency and prompt accuracy. Rolling out in Gemini app, Search & Ads Have you tried it yet? In my opinion, the Chinese still do some things better blog.google/innovation-and-a…

Google drops Gemini 3.1 Pro – built for complex, high-intelligence tasks. Standouts: - Animated SVGs interactive 3D from text - Full functional websites & dashboards - Stronger agentic workflows & reasoning Live in preview. New reasoning leader? #AI #Gemini #Google #AgenticAI blog.google/innovation-and-a…

x.com/Reuters/status/2024401… Anthropic just democratized frontier AI: Claude Sonnet 4.6 now matches or beats flagship Opus 4.6 on coding, agent workflows, finance, and enterprise tasks – at 1/5th the price! Standout upgrades: 1M token context (beta) = entire multi-day projects in one window. Sharper agentic brain: smarter planning, fewer hallucinations, cleaner execution. Blind tests: users preferred Sonnet 4.6 over Opus 4.5 in 59% of head-to-heads for speed and quality. Rolled out as default for Free & Pro users on claude.ai full API/cloud support. The gap between “elite” and “everyday” AI just collapsed. Enterprise adoption is about to explode. Anthropic making top-tier AI affordable for everyone – biggest shift of 2026? #AI #Anthropic #Claude #Sonnet46 #AIAgents anthropic.com/news/claude-so…

Apple Accelerates AI Wearables Push: Ramping up three new devices—smart glasses, a wearable pendant (necklace/pin), and camera-equipped AirPods—to supercharge Siri with real-world visual awareness! Highlights: • All feature built-in cameras iPhone tethering for context-aware actions. • Smart glasses (premium with high-res vision tech) → production late 2026, launch likely 2027. • Pendant and upgraded AirPods → potential releases as soon as this year or next. • Direct response to Meta’s Ray-Bans, Google’s efforts, and OpenAI’s hardware ambitions. Apple betting big on ambient, always-on AI to lock in ecosystem loyalty. The future of wearables? Ready for Siri that actually sees what you see? 🔥 bloomberg.com/news/articles/… #AI #Apple #Wearables #Siri #AIHardware

OpenAI ramps up security: ChatGPT now features "Lockdown Mode" to shield high-stakes users from sneaky attacks like prompt injections that could expose sensitive info. Key features: - Toggle-on option that predictably shuts down exploitable tools. - Web browsing sticks to pre-cached data only – no external calls from OpenAI's servers. - Workspace admins control it, with options to greenlight safe apps/actions. - Plus, "Elevated Risk" tags popping up in ChatGPT, Atlas, and Codex to highlight potential vulnerabilities. This move signals the era of AI agents is here, where models handle web access and tasks—making ironclad defenses essential to prevent mishaps. Security boost or overkill? How might this change your AI workflow? openai.com/index/introducing…

The Pentagon is close to cutting ties with Anthropic, labeling it a "supply chain risk" (reserved for foreign adversaries) over restrictions on Claude's military use. Details: Would force all U.S. defense contractors to drop Anthropic, crippling its business. Officials demand "all lawful purposes," but Anthropic resists broad permissions for spying on Americans or autonomous weapons. Claude is the only AI on Pentagon classified systems, reportedly used via Palantir to capture Maduro in January. Company open to loosening but holding firm on ethics. Why it matters: Friction between AI guardrails and military demands highlights warnings about unchecked AI in warfare. Claude in the military: Boon or threat? Thoughts? #AI #Anthropic #Pentagon #Claude #AIethics #NationalSecurity axios.com/2026/02/15/claude-…

ByteDance just dropped Seed 2.0 – a new AI model family that matches or beats GPT-5.2 and Gemini 3 Pro across dozens of math, reasoning & vision benchmarks… at nearly 1/10th the price! 💥 Key facts: • Seed 2.0 Pro: $0.47/M input tokens (vs $1.75/M GPT-5.2 and $5/M Gemini 3 Pro) • Built for real-world agentic tasks – demo: autonomously completes 96-step CAD modeling workflows • Follows viral Seedance 2.0 video model that sparked Hollywood IP controversy • Live now on Doubao app (Expert Mode) API (limited outside China) China is shaking up the AI leaderboard with bargain prices. Who’s next? seed.bytedance.com/en/seed2 #AI #ByteDance #Seed2 #GenerativeAI #LLM

OpenAI's GPT-5.2 Pro just discovered a new mathematical formula in theoretical physics—proving a long-assumed zero amplitude in gluon interactions is actually nonzero in the half-collinear regime! It conjectured, simplified, and even formally proved it autonomously. Verified by experts like Nima Arkani-Hamed and Nathaniel Craig. First original AI contribution to physics? 🤯 Quotes: "Exciting to see strikingly simple expressions... looking forward to a 'simple formula pattern recognition' tool." - Arkani-Hamed Is AI redefining scientific discovery?

But when she puts on makeup, she can be anyone, even a man - it's very versatile 😅

OpenAI just launched GPT-5.3-Codex-Spark – and it runs on Cerebras, not Nvidia. The new model is optimized for speed, generating 1,000 tokens per second, making it OpenAI’s first product powered outside its traditional Nvidia stack. Spark trades some raw intelligence for latency, underperforming the full 5.3-Codex on SWE-Bench Pro and Terminal-Bench – but completing tasks far faster. The idea: Spark handles rapid interactive edits, while the full Codex runs longer autonomous jobs in the background. The release follows OpenAI’s $10B deal with Cerebras and additional agreements with AMD and Broadcom – a clear push to diversify compute. Spark is rolling out as a research preview for ChatGPT Pro users, with limited API access for select enterprise partners. Speed has been Codex’s biggest weakness. This update tackles that – and makes OpenAI’s chip strategy tangible. openai.com/index/introducing…

Google just dropped a major upgrade to Gemini 3’s Deep Think mode – and the benchmark numbers are aggressive. The model scored 84.6% on ARC-AGI-2, well ahead of Opus 4.6 (68.8%) and GPT-5.2 (52.9%), and set a new high of 48.4% on Humanity’s Last Exam. It also posted gold-medal level results on the 2025 Physics and Chemistry Olympiads, and reached 3,455 Elo on Codeforces – nearly 1,000 points above Opus 4.6. Alongside the upgrade, Google introduced Aletheia, a math agent designed to autonomously solve open problems and verify proofs. Deep Think is live for AI Ultra subscribers, with API access rolling out to researchers via early access. After much of 2026’s attention shifted toward Anthropic and OpenAI, Google is reminding the field it’s still very much in the race – especially in math and science reasoning.

Briefly at Davos: AI leaders set a timeline. @AnthropicAI Dario Amodei said we’re 6–12 months from AI handling most software engineering and warned on U.S. chip sales to China. @Google sees pressure on junior hiring. @Microsoft says slow movers will be overtaken.

You can now code with Codex from your phone Set it up in Cursor, push the repo to GitHub, and Codex runs and reviews code in the cloud – including automatic PR reviews A laptop is no longer required

ChatGPT with ads – now official OpenAI is testing targeted Sponsored Recommendations for free and Go users. No politics or health, and a promise that ads won’t influence answers Altman once called ads a “last resort” IPO pressure seems to be changing the calculus

New Teslas drive themselves off the line at Giga Berlin Autonomously navigating the factory premises, they stop by the on-site Supercharger, then park in the outbound lot

First official Light Show (Ready For Assault Multi-Vehicle) at the Tesla Diner! 15 Cybertrucks. Thanks to everyone who came out to join in on the fun and make this happen! Thanks to @FthePump1 and @CyberOdysseyUSA for putting the word out and the Tesla staff and security for helping with the coordination. Thanks to @chattybird0306 for helping with the logistics. And thanks to @TeslaCharging for creating a special place for the Tesla Community! We love you all! Please tag yourself if we missed your handle. @jehazel75 @chattybird0306 @FthePump1 @SoCalEVs @CyberOdysseyUSA @cyberoffroad @BLKMDL3 @brand0n @aaronjcash @dogetothemoon @techiedani @ResilientRye @ItsKimJava @misterjr24 @elonmusk @TeslaCharging @woodhaus2 @cybertruck @Tesla_Optimus @TeslaOwnersMS

ATARI is re-releasing the 2600 console in a limited PAC-MAN edition for the franchise’s 45th anniversary 👾 The bright yellow console ships with two games: the original 1982 PAC-MAN port for the 2600 and a modern remake. It supports vintage Atari 2600/7800 cartridges, includes a wireless CX-40 joystick — and yes, it can play chess too. $170. Launching late October. atari.com/products/pac-man-a…

Found a MEGACONVERTER that runs locally and supports over 1,000 file formats 🧩 • Fully offline — nothing touches the internet • Handles any file type, from docs to media to data dumps • Batch conversion saves hours • Password protection & multi-user support built in • 100% free A must-have tool for anyone dealing with files 👉github.com/C4illin/ConvertX