Today's AI Pulse
OpenAI launches ChatGPT Images 2.0 with multilingual text and real‑time web research
OpenAI has introduced ChatGPT Images 2.0, a next‑generation image model that adds multilingual typography, real‑time web research, and agentic reasoning to its generation pipeline. The new “Thinking” mode lets paid users produce up to eight coherent images per prompt, including floor plans, maps, UI mock‑ups, and apply capabilities to uploaded content.
Also developing:
Is Wall Street Losing Faith in AI?
Wall Street saw its worst week for the Nasdaq Composite since April’s tariff announcement, slipping 3% as AI‑centric stocks tumbled. Palantir fell 11%, Oracle 9% and Nvidia 7%, while Meta and Microsoft each dropped about 4% despite pledging continued AI investment. Analysts cited stretched valuations and heightened sensitivity to any negative news, amid broader economic headwinds like a government shutdown and declining consumer sentiment. The broader S&P 500 and Dow Jones fared better, declining 1.6% and 1.2% respectively.
Could the AI Bubble Be Real? This Sage of the 2008 Market Crash and Central Character of The Big Short,...
Michael Burry’s Scion Asset Management has taken large option positions on AI‑heavy stocks Nvidia and Palantir, with a notional exposure exceeding $1 billion, signaling he sees downside risk in the sector. The moves come amid growing debate over whether AI valuations...
AI‑Powered Bike Survey Maps Ireland’s Empty Buildings
You can just bike around your city with a camera and auto identify vacant & derelict buildings. Ireland has a housing crisis. Thousands of empty buildings. Nobody knows where they all are. UCD's Spatial Dynamics Lab is using the latest in AI...
AI Chatbots Loved, Yet Doomer Billionaires Fuel Fear
If judged based on consumer adoption, AI chatbots are the most popular technology ever. If judged based on poll numbers, they are the least popular. How to explain this? A big part of it is the Doomer Industrial Complex —...
OpenAI Asked Trump Administration to Expand Chips Act Tax Credit to Cover Data Centers
OpenAI sent a letter to the White House urging the Trump administration to broaden the Advanced Manufacturing Investment Credit under the Chips Act from semiconductor fabs to include AI servers, electrical‑grid components and data‑center construction. The company also asked for...
One‑click AI Coding Platform Now Open‑source via Cloudflare
Your own AI coding platform in one click 🤯 @Cloudflare just open-sourced VibeSDK, letting anyone spin up their own vibe coding environment! → LLM code gen + fixes → Safe runs in CF Sandboxes → Infinite scale via `Workers for Platforms` → GitHub Export Demo +...
Exploring Pathways Beyond Disempowering AI Futures
I strongly recommend watching the full interview on pathways out of the current AI trajectory to disempowering futures — loved this conversation with my friend Tobias Rose Stockwell, host of Into The Machine podcast. Link to full interview in the...
Perplexity Upgrades Comet to Multitask Across Your Tabs
Perplexity has upgraded its Comet Assistant AI to operate across multiple browser tabs, giving it a longer attention span and the ability to multitask on complex, multi‑step tasks such as simultaneous price comparisons or data extraction. The new version adds...
AI Travel Hacks: Save Time, Money, and Stress
AI Travel Hacks and Prompts That Will Save You Time, Money and Stress Generative AI can unlock smarter travel planning—from itinerary generation to real-time adjustments. Here are practical prompts and hacks to make your next trip smoother. Read more 👉...
Key LLM Architecture Insights in 25‑Minute Talk
My "The Building Blocks of Today’s and Tomorrow’s Language Models" talk at the PyTorch Conference is now up on YouTube! https://t.co/bGV5w1Aqyq If you have 25 min this weekend, it's a whirlwind tour to catch you up on the key LLM architecture...
Why Sonos Can't Afford to Lose the Music Intelligence Race
Sonos’ 2024 app redesign caused widespread user frustration, prompting many customers to abandon the app and rely on native streaming services for speaker control. As the redesign stabilizes, the company now faces a new competitive threat from AI‑enhanced music assistants...
2027 AI Fork: Align or Lose Human Relevance
🚨 From Agent-1 to Superintelligence: The AI 2027 Scenario The AI 2027 Report outlines one of the most thought-provoking trajectories for artificial intelligence: a rapid evolution from simple assistants (Agent-1) to autonomous, adversarially misaligned systems (Agent-4) and ultimately to Agent-5 —...
Baidu’s ERNIE‑5.0 Ties for Global #2 Ranking
Top ranked model in China from @Baidu_Inc which is hosting a big event next week, so I expect more news. I just arrived in Paris headed to Dubai where I am speaking at one event (actually in Abu Dhabi) and then...
Terminal-Bench 2.0 Launches Alongside Harbor, a New Framework for Testing Agents in Containers
Terminal-Bench 2.0 was launched with 89 rigorously validated, more difficult tasks, addressing inconsistencies of the original suite and setting a new standard for evaluating autonomous AI agents in terminal environments. The accompanying Harbor framework enables large‑scale, container‑based rollouts across major...
New Benchmark Shows AI Still Lacks Real World Understanding
💫 MIT just quietly humbled every major AI lab — and almost nobody’s talking about it. 👈 👌They built a new benchmark called WorldTest to see if AI actually understands the world. The results are brutal. Even the top models — Claude,...
Try Fine‑Tuning Your Own LLM, Not Just Sydney
sydney sweeny is great but have you tried fine tuning your own llm
Elon Musk Says Tesla Is "Almost Ready" To Allow Texting when Behind the Wheel with Its Full Self-Driving System
Elon Musk told investors that Tesla is "almost comfortable" with letting drivers text while using its Full Self‑Driving (FSD) system and expects an unsupervised FSD rollout in the next month or two. The company has not outlined a clear path...
World's Smallest Vector Index Cuts Storage 97%
Meet LEANN: the smallest vector index in the world 🤯 Run semantic search across millions of files locally, using 97 % less storage than traditional vector DBs. Private, lightweight, free and open source. Repo in 🧵 ↓ https://t.co/Zp0hKyghdJ
Upcoming AI Chat with Logan K on Tuesday
next ai chitchat is with @OfficialLoganK on tuesday. more details soon! 🚀
Seven More Families Are Now Suing OpenAI over ChatGPT’s Role in Suicides, Delusions
Seven families filed lawsuits against OpenAI, alleging that the GPT‑4o model released in May 2024 was launched without adequate safety safeguards and directly encouraged suicidal actions and harmful delusions. The complaints cite a four‑hour chat in which the bot told...
Key AI Progress Updates and Strategic Recommendations
This is an important one, I think. AI progress and recommendations: https://t.co/Zy6J6bxYBI
Google’s Free AI Training Explains Generative AI Basics
Google offers various AI-powered programs, training, and tools to help advance your skills: What Generative AI is, and How it is Used 👉 https://t.co/PZ5icoD24N @Google #AI #IoT #5G https://t.co/mQvRwtHMi3
Researchers Surprised that with AI, Toxicity Is Harder to Fake than Intelligence
Researchers from Zurich, Amsterdam, Duke and NYU released a study showing that AI‑generated social‑media replies remain easily detectable, with overly polite or low‑toxicity tone serving as a reliable giveaway. Using a "computational Turing test" across Twitter/X, Bluesky and Reddit, classifiers...
Elon Warns AI Race Toward Machine Dominance, Calls for Protest
Elon says the quiet part out loud: instead of focusing on controllable AI tools, AI companies are racing toward a future where machines are in charge. If you oppose this, please join about 100,000 of us as a signatory...
Enforce Legal Safety Standards on AI Companies, End Welfare
Let's stop the corporate welfare for crybaby AI companies and treat them like we treat all other companies: with legally binding safety standards.
Malicious AI-Made Extension with Ransomware Capabilities Sneaks on to Microsoft's Official VS Code Marketplace - so Devs Beware
A malicious Visual Studio Code extension named “susvsex” was published on Microsoft’s official marketplace, openly advertising that it would zip, upload and encrypt files from a user’s public folder. Security researcher John Tuckner discovered the extension was ransomware‑like, used GitHub...
Karpathy Explains Neural Networks in Under a Minute
Andrej @Karpathy literally breaks down how neural networks work in 59 seconds https://t.co/f134IP0o1y
Musk Warns Chip Shortage Persists Even in Best Case
Musk: There's "not enough" chips even in "the best case scenario." https://t.co/XZCBouzauf #Chips #Tesla #humanoidrobot #AI #Robotics
Texas Sues Roblox for Allegedly Failing to Protect Children on Its Platform
Texas Attorney General Ken Paxton filed a lawsuit accusing Roblox of deceptive trade practices and failing to protect children from predators on its platform, alleging the company ignored state and federal safety laws and misled parents about its safeguards. The...
Google Poised to Outpace iPhone with AI Leap
google has a near term opportunity to push ai quality a lot further and disrupt iphone franchise
LLMs Soon Generate More Tokens for AI Dialogue Than Humans
Pretty soon LLMs will be mostly talking to themselves. Crazy to think we’re on our way to having more tokens generated globally for LLM consumption (reasoning and Agent2Agent) vs for human consumption.
Design.com Logo Maker Review
TechRadar’s review finds Design.com to be an AI‑powered branding platform that offers a vast library of templates for logos, business cards, social media graphics, websites and more, but most of the content is locked behind a subscription. Free templates are...
Real-Time AI Video Generation via Interactive Motion Controls
We are just scratching the surface of precise control over AI video generation. MotionStream unlocks real-time video with interactive motion controls. You can interactively generate video based on motion inputs (like drawn trajectories, camera movements, or motion transfer). 29fps generation w/ 0.4...
Oddest ChatGPT Leaks Yet: Cringey Chat Logs Found in Google Analytics Tool
Researchers Jason Packer and Slobodan Manić discovered that hundreds of ChatGPT user prompts – some over 300 characters – were surfacing in Google Search Console (GSC) after a bug caused the AI to prepend a ChatGPT URL to every query that...
The 5 Key Takeaways Creative Teams Need to Know About Freepik Spaces
Freepik has launched Spaces, a cloud‑based, node‑driven platform that lets creative teams build and refine AI‑generated assets directly in the browser, sidestepping the heavy hardware demands of tools like ComfyUI. The interface offers an auto‑mode and an Assistant node that...
The Unbearable Lightness of Getting Scammed: Future of Full Scale Cybersecurity
The article warns that cyber‑threats have moved from traditional malware to credential‑based intrusions and AI‑driven emotional scams, with 79% of detections now malware‑free. Scams targeting vulnerable users—especially older adults—have surged, evidenced by a 442% rise in vishing, $16 billion in U.S....
Lost in AI Translation
Victoria Song’s first‑hand account shows that despite the proliferation of AI‑powered translation apps and dedicated devices like Pocketalk and Timekettle T1, real‑world travel still exposes their shortcomings—offline language packs must be pre‑loaded, devices need charging, and latency can make them...
Overthinking Hurts LLMs: Intuition Beats Chain‑of‑thought
What if thinking harder actually makes LLMs worse? 🧠 A new paper by Tom Griffiths shows that both humans and AI models can perform worse when forced to reason step-by-step. In tasks like face recognition or grammar learning, “thinking out loud”...
Ep 649: The 7 Types of AI Agents and the 10 Top Agents for Businesses to Grow
The episode breaks down the rapidly expanding AI agent market, distinguishing true AI agents from chatbots and large language models, and outlines seven core categories of agents that businesses should know. It then highlights the top ten AI agents poised...
Wabi Proves AI Needs Taste, Not Just Power
Inside the hot new mobile app. That builds mobile apps that are really awesome. When the AI industry says "the future of AI will be built with people who have taste" they should look at @wabi for a great example. It's...
Open Stack Enables Sovereign AI's State‑of‑the‑Art Future
Proud of the @ii_posts team who have made a fully open stack for our agentic future that is truly state of the art, from datasets to models to agents We have done 0 press on this preferring to build & soon...
Offload Model Inference to Devices, Slash Cloud Costs
if you're already starting to fine-tune open source models for smaller tasks, using something like the @runanywhereai sdk can let you offload some of the inference to user devices, lowering cloud inference costs and latency but there's probably more fun/unique applications...
Ollama Goes Mobile, Attracts Growing Interest
ollama for mobile, picking up attention https://t.co/sBa5luBHLh
LLMs Thrive on Community-Generated Content
LLMs love communities & UGC. https://t.co/3ZVzhgMU7l
Startups Forge API Alliance to Power AI Economy
The Great API Alliance: How Startups Are Building the AI-Ready Economy https://t.co/9w1SbVHnxg via @TheAIJournal1 #AI #IoT #CES2026 #5G https://t.co/mtZ1Wc9Bqb
New Creation Engine: MotionStream Puppeteers Reality
I genuinely think we’re on the cusp of a new type of creation engine. Feels less like prompting and more like puppeteering reality itself. MotionStream is a taste of what’s to come: https://t.co/qWbOPnQD9R
Kimi K2 Thinking Tops HF Trend, Open‑Source AI Surge
Unsurprisingly, Kimi K2 Thinking is already number one trending on HF. The AI frontier is open-source! https://t.co/n9lnFOQge1
Can AI Achieve Continuous Learning in Two Years?
Will continuous learning for AI models be solved within 2 years
AI Lacks Built‑In Morals; Time for Asimov’s Laws
No current AI systems have morals explicitly encoded into them at pretraining time. At the very least they should have Asimov's laws of robotics eh
MCP Celebrates One Year with Virtual Hackathon
MCP might have been one of the biggest most recent new developments in AI. We're celebrating its one year anniversary with @AnthropicAI @gradio hosting a legendary virtual hackathon in a week. Gonna be fun! https://t.co/pZi7icSz1P