In a memo, Alexandr Wang wrote that "superintelligence is coming." In order to "take it seriously," the company is making major changes. MSL's research will be led by Shengjia Zhao. Yann LeCun will report to Wang. (great scoop from @CharlesRollet1) https://t.co/woKclcDcSF

Source: https://t.co/n5pSzPMYZc

Makes you think about how we approach these technologies. Source: https://t.co/kvljKHGe7P

our rights over our own images and voices?” Source: https://t.co/kc9P8xxgrb

🧬@Latent_Labs has a vision of “computational mastery of biology,” using #AWS solutions & #generativeAI to make the science “programmable.” Discover how: 👉 https://t.co/UfzqTWH3IC https://t.co/wqyc64HjiJ

The Fitbit App Is Turning Into an AI-Powered Personal Health Coach https://t.co/6zoIgU4w6X @wired

The Fitbit App Is Turning Into an AI-Powered Personal Health Coach https://t.co/6zoIgU4w6X @wired

Edit images in Google Photos by simply asking https://t.co/HvAN17Crwl @google https://t.co/9MXKbrBQBw

Edit images in Google Photos by simply asking https://t.co/HvAN17Crwl @google https://t.co/9MXKbrBQBw

I Saw the Future of AI Film and It Was Empty https://t.co/r0krwJvCz6 @johnsemley3000 @wired

I Saw the Future of AI Film and It Was Empty https://t.co/r0krwJvCz6 @johnsemley3000 @wired

What is ‘AI psychosis’ and how can ChatGPT affect your mental health? https://t.co/KyML4i383a @washingtonpost

What is ‘AI psychosis’ and how can ChatGPT affect your mental health? https://t.co/KyML4i383a @washingtonpost

How AI-enhanced hackers are stealing billions https://t.co/ZVaPsJTwuX @theeconomist

How AI-enhanced hackers are stealing billions https://t.co/ZVaPsJTwuX @theeconomist

Meta’s $29 Billion Deal Exemplifies the High Financial Stakes of AI https://t.co/2rXvZP4Tu7 @business

Insightful Pulse of Fintech – UK perspective https://t.co/tSHmQZJUAH @kpmguk @kpmg https://t.co/IyOwS6th8Y

Chris Cuomo declares the Democrat Party is DEAD 💀 Cuomo tells me elitism, open borders, socialism and defund the police killed the modern Democrat Party. Cuomo says the Democratic Party that his father was apart of “no longer exists” and he “doesn’t know why” his brother still registers as a Democrat. “My brother’s a Democrat. I don’t know why, but he is. My father was a Democrat. But his party doesn’t exist anymore.” “Now it's the left arguing for a cultural elite. This is how you will talk about people, Benny. This is how you will raise your kids. This is what you will tell them is okay. This is what you will say and not say.” “There are socialistic aspects to our constitutional republic within a capitalistic model that work. They work for the collective. But it's a capitalistic society. No Democrat ever argued for anything else. No Democrat would have argued for open borders. None of this. My father would have done none of this.” “And while I had disagreements with my father about different issues, I knew what principles were guiding him.” Wow. Well said @ChrisCuomo

AI stocks are cracking. Nvidia is sliding. Palantir just dropped 10% in a day. https://t.co/9drkpPNjGl, Salesforce, Adobe and others are falling fast. For 18 months, AI carried the Nasdaq and S&P. Now it’s dragging them down. So what’s really happening? 1.The Trigger •Profit-taking: August is historically weak. •Narrative shift: “AGI is coming” → “GPT-5 flopped” → “AI is a bubble.” •MIT study: 95% of enterprise GenAI pilots delivered zero ROI. Only 5% created measurable business impact. That stat alone spooked markets. 2.Sam Altman’s Warning Even OpenAI’s CEO admits: “AI is in a bubble.” Why now? •To cool down competitors raising at inflated valuations. •To prepare Microsoft and regulators for turbulence. •To reset expectations before hype collapses. This is strategy, not altruism. 3.Short Term Reality •Expect more volatility and repricing. •Short sellers will target inflated names. •90% of AI startups won’t survive. •Public stocks with weak fundamentals will get punished. The shakeout has begun. 4.Long Term Reality This is not the dot-com bust replayed. •AI is foundational, not optional. •Every major sector such as healthcare, finance, defense, energy is embedding AI into core operations. •Like the internet in the 1990s, there is a “productivity paradox”: investment first, payoff later. •95% may fail, but the 5% that survive will become the Amazons of AI. 5.The Takeaway Yes, air is coming out. Yes, valuations will be tested. But AI itself is not a bubble. The bubble is in the copycats, hype decks, and inflated multiples. The technology is inevitable. The long-term trajectory is unstoppable. POV: Short term = pain, volatility, shakeout. Long term = trillions in value creation and a handful of new global giants. The future belongs to AI.

No doubt, @WorkOS is crushing it on all things community… MCP Night, now this! https://t.co/TsTZboDJdT

New blog post by @AmanGokrani: Everyone says Claude Code "just works" like magic. He proxied its API calls to see what's happening. The secret? It's riddled with <system-reminder> tags that never let it forget what it's doing. (1/6) [🔗 link in final post with system prompt] https://t.co/Ne3X4DivzS

RAG is not Memory for AI Agents. 5 AI memory engines to build agents that maintain long-term context and learn continuously: (Last 2 released just this month) 1. Zep builds and queries temporally-aware knowledge graphs that evolve with every interaction. 100% Opensource. https://t.co/8DaY35Xjiq

One of the quickest ways to start playing with a good local LLM on macOS (if you have ~12GB of free disk space and RAM) - using llama-server and gpt-oss-20b: brew install llama.cpp llama-server -hf ggml-org/gpt-oss-20b-GGUF \ --ctx-size 0 --jinja -ub 2048 -b 2048 -ngl 99 -fa https://t.co/xsvruU8hDD

Back in 2017 my book had an entire chapter on generative AI, including language modeling and image generation. I wrote that content in 2016 and early 2017. This was some of the earliest textbook content that covered generative AI. All the way back in 2014 I was convinced that AI would one day become a major source of cultural content creation -- which was a completely outlandish position at the time.

One of the best illustrations of Moore’s Law: on the left is the Apollo Guidance Computer, which took humans to the moon. To the right is an Anker USB-C charger that can charge 2 phones. The charger has 48x the clock speed & 1.8x more memory than the AGC! https://t.co/Ml5mYSCi5n https://t.co/4nN9UnNtN6

@lucataco93 @bella_rwa . @KitDelIntent remove makeup from this person https://t.co/q1WQDg3XPU

If you’re in AI pivot to snake pets @CoinfestAsia https://t.co/jAXV7cw5MM

@marouen19 @lucataco93 @bella_rwa Makeup removed. Here’s a hyper-realistic version with only natural skin and features. All age and facial detail kept unchanged; no stylization, just facts. (Let me know if a stricter before/after split is needed—this was a straight facial clean-up.) https://t.co/t6zJN6ocEk

This is how you enter the shadow realm https://t.co/pbPFL3cJtd

@kalomaze Also remember the one time Andrej roasted the W&B social media team cuz they originally didn't let me in lol https://t.co/NVVbbCkTRM

DuPO: Enabling Reliable LLM Self-Verification via Dual Preference Optimization "We present DuPO, a dual learning-based preference optimization framework that generates annotation-free feedback via a generalized duality" "DuPO decomposes a primal task’s input into known and unknown components, then constructs its dual task to reconstruct the unknown part using the primal output and known information (e.g., reversing math solutions to recover hidden variables"

Your Reward Function for RL is Your Best PRM for Search: Unifying RL and Search-Based TTS "we leverage adversarial inverse reinforcement learning (AIRL) combined with group relative policy optimization (GRPO) to learn a dense, dynamic PRM directly from correct reasoning traces, entirely eliminating the need for labeled intermediate process data. At inference, the resulting PRM simultaneously serves as the critic for RL rollouts and as a heuristic to effectively guide search procedure"