971972
🤖 The #1 AI news source! We cover the latest artificial intelligence breakthroughs and emerging trends. Manager: @rational
SpaceX has finalized a major agreement with open-source AI startup Reflection, granting the company access to Nvidia GB300 GPUs through a compute lease valued at $150 million per month, beginning in July 2026. Total payments could reach approximately $6.3 billion if the contract is maintained until 2029.
Either SpaceX or Reflection can terminate the deal with 90 days’ notice after the initial three months. This marks a shift for SpaceX, which is now positioning itself as a GPU cloud provider, leveraging its Colossus infrastructure for external clients.
Reflection’s business model centers on building and customizing frontier open-weight AI models for government and enterprise use, requiring substantial compute capacity to train these foundational models. Recent changes by other AI providers have increased demand for open-access model solutions, supporting Reflection’s approach.
📰 @aipost
🧠 Sam Altman just dropped a brutal observation about AI.
He said some of the smartest scientists in the field were also the people who slowed it down the most.
Not because they lacked talent. Because they were too sure they were right.
For years, respected experts insisted scaling wouldn’t work. Bigger models wouldn’t lead to major breakthroughs. The approach had hit a wall.
Then reality showed up.
The models kept getting better. New capabilities kept emerging. The predictions kept failing. But many critics didn’t change their minds.
Altman says that’s what happens when a belief becomes part of your identity.
Once your reputation is built around a particular view, updating your opinion starts to feel like admitting defeat. And the smarter you are, the better you become at defending a position long after the evidence has moved on.
He even compared it to a kind of insanity: seeing the data change while repeating the same conclusion.
But his warning wasn’t aimed only at the skeptics. It was also aimed at the people winning today.
Because the moment any belief becomes who you are, instead of something you hold, you lose the ability to update.
And in a field moving this fast, the ability to change your mind may be the most valuable skill of all.
@aipost 🏴
Day 1 of vibecoding 😂
@aipost 🏴
🔥Sakana AI just launched a model that doesn’t act like a model.
Meet Fugu: a single API that secretly runs an entire team of AI agents behind the scenes.
Instead of answering everything itself, Fugu can:
• Pick the best model for the job
• Delegate tasks to specialist agents
• Verify results
• Combine everything into one final answer
The wild part? Fugu can even call copies of itself recursively. To developers, it looks like one model. Under the hood, it’s an AI manager coordinating an entire workforce of AIs.
Sakana says its flagship Fugu Ultra performs alongside frontier models like Fable and Mythos on demanding reasoning, science, and engineering benchmarks without the export-control headaches tied to some cutting-edge systems.
Think of it as the difference between hiring one genius and hiring an entire company through a single email address.
The age of “one model = one brain” may be ending. The future might be AI teams masquerading as a single model.
Source.
@aipost 🏴
⚠️Silicon Valley’s richest people are betting that aging is a bug and AI might be the fix.
Anthropic CEO Dario Amodei thinks AI could compress 100 years of biological progress into just 5–10 years. Sam Altman has personally invested $180 million into Retro Biosciences, a startup trying to extend healthy human lifespan.
Peter Thiel has spent years funding longevity research, backed the Methuselah Foundation, experimented with human growth hormone, and has openly discussed young blood transfusions as a potential anti-aging therapy.
Jeff Bezos poured $3 billion into Altos Labs, making it one of the most heavily funded biotech startups ever. He has also backed Unity Biotechnology, Sana Biotechnology, and Denali Therapeutics.
Coinbase founder Brian Armstrong argues aging is the world’s biggest killer, claiming it causes more than 100,000 deaths every day. He launched NewLimit to develop treatments that could slow or reverse the process.
Larry Ellison has funded anti-aging research for more than two decades.
Even Elon Musk, despite warning about the risks of leaders living forever, says longevity is an “extremely solvable” problem.
The biggest race in tech may no longer be building AI, it may be using AI to make aging optional.
@aipost 🏴
Amazon may be making its biggest move yet against Nvidia.
The company is reportedly exploring selling its Trainium AI chips directly to data centers and enterprises, instead of keeping them exclusive to AWS.
That would be a major shift. Today, companies need Nvidia GPUs or cloud access to train AI models. Tomorrow, they could buy Amazon’s chips and run them on their own infrastructure.
The timing is notable: OpenAI, Anthropic, and Uber are already using Trainium through AWS, and Amazon says demand for its AI chips is surging.
The bigger threat to Nvidia? Amazon isn’t alone. Google is also starting to offer its TPUs to external customers.
For years, hyperscalers were Nvidia’s biggest customers. Now they’re becoming its biggest competitors.
@aipost 🏴
Amazon just dropped its Sam Altman movie and the timing is raising eyebrows.
Amazon MGM has reportedly walked away from Artificial, Luca Guadagnino’s film about the dramatic 2023 OpenAI boardroom coup that briefly saw Sam Altman fired and then reinstated as CEO.
The twist? Amazon recently deepened its partnership with OpenAI, committing tens of billions of dollars to the AI company. Soon after, the studio decided the film would be “better served” by another distributor.
The movie stars Andrew Garfield as Altman, Ike Barinholtz as Elon Musk, and Yura Borisov as OpenAI co-founder Ilya Sutskever. Reports suggest the film portrays several AI leaders in a less-than-flattering light.
So a film about one of Silicon Valley’s biggest power struggles is now looking for a new home, right after the studio behind it became one of OpenAI’s biggest partners.
@aipost 🏴
Within 5 years time, Claude thinks we will not know what the truth is anymore.
@aipost 🏴
🔥Genesis AI introduced Eno, its first general-purpose robot, with availability planned for the fourth quarter of this year.
Unlike traditional humanoid robots, Eno is not designed to closely imitate the human body. Its form is built around practical movement, physical intelligence, and the ability to assist across different environments.
@aipost 🏴
Singapore leads in AI use at work, report finds
Microsoft’s 2026 Work Trend Index reveals that Singapore’s workforce is advancing more quickly in AI adoption than global averages.
According to the report, 66% of AI users in Singapore say they are producing work that would not have been possible for them a year ago. By comparison, the global figure is 58%.
This proportion increases further among ‘Frontier Professionals’—the company’s term for highly advanced AI users—with 82% of this group in Singapore reporting new work capabilities due to AI tools.
📰 @aipost
Scientists administer first gene therapy to rejuvenate human cells
For the first time, a patient has received a gene-based treatment intended to reverse characteristics of cellular aging, returning old cells to a more youthful state without converting them into stem cells.
The therapy is part of a clinical study led by Life Biosciences that focuses on individuals with glaucoma. The objective is to regenerate optic nerve cells damaged by the disease.
A key question remains whether cellular rejuvenation techniques can be applied safely in humans. This trial represents an initial step in exploring the potential for therapeutic reversal of age-related changes in biology.
📰 @aipost
AI just took my damn job.
@aipost 🏴
⚠️Your Claude account may soon need a passport and a selfie to log in.
Starting July 8, Anthropic will begin “selectively” requiring users to verify their age and identity with a government-issued passport and a selfie before accessing Claude.
The company says the move is designed to keep its services “safe and secure” and to meet evolving regulatory requirements.
The new verification rules will apply to Free, Pro, and Max subscribers. Business plans and API users will not be affected.
For now, Anthropic says the checks will be rolled out selectively, though anyone who has watched tech platforms introduce “optional” verification before knows how these things can evolve.
Source.
@aipost 🏴
Researchers publish atlas of human cellular senescence
Scientists have released the first detailed map showing where senescent— or “zombie”— cells are found in the human body as we age. This atlas identifies the specific tissues where these non-dividing, persistent cells accumulate, and examines their differences and how they interact with the local environment.
Senescent cells remain alive but stop dividing. They can build up in tissues and emit inflammatory signals linked to damage of nearby cells. This process is associated with conditions like arthritis, cancer, Alzheimer’s disease, and general immune decline.
The publication of this atlas creates a foundation for targeting specific senescent cell types, aiming for precision therapies focused on aging and age-related diseases.
📰 @aipost
❗️Vivek, a researcher at Anthropic, published an article on how to be good at research.
Here are the 7 core principles as prompt instructions:
1. "Restate the problem in your own words and define what a successful answer looks like before responding."
2. "Predict the most common answer to this question. Then tell me what it gets wrong or misses."
3. "Prioritize foundational sources over recent or popular ones. Older proven frameworks beat trending takes."
4. "Show your reasoning chain. State each assumption and flag where your logic is weakest."
5. "Start with the smallest possible version of this problem. Solve that first, then add complexity one piece at a time."
6. "After answering, list the three strongest objections to your own response and tell me which one is most valid."
7. "Draw from at least one adjacent field outside the obvious domain for this question."
🗣Demis Hassabis believes AI could revolutionize medicine within the next few years.
- AI already solved the 50-year protein folding challenge with AlphaFold.
- Future AI systems could design new drugs automatically.
- Drug discovery could shrink from 10 years to months, weeks, or even days.
- Most experiments may happen in simulations before human validation.
- Personalized medicines tailored to individuals could become possible.
- Demis believes AI could bring all diseases within reach of treatment.
Future of healthcare could be extraordinary!
@aipost 🏴
🗣Dario Amodei, anthropic's CEO, just answered the question everyone is asking. Should you still learn to code?:
1. Coding is going away first. the AI models are doing it already. The broader task of software engineering takes longer but that's going too. If you're learning to code purely for job security, you're learning the wrong thing.
2. Even at 5% of the task you're still valuable. If AI does 95% and you do 5%, you become 20 times more productive. Comparative advantage is surprisingly powerful even when the gap is massive.
3. The professions with the most runway are human-centered ones. Things that mix people, the physical world, and analytical skills together. He uses the radiologist example. The doctor who understands patients and context, not just reads scans.
4. Critical thinking might be the most important skill of the next decade. When AI can generate anything, the ability to tell what's real from what's fake becomes rare and valuable. You don't want false beliefs. You don't want to get scammed. That’s his actual advice to a 25 year old.
5. AI can make you stupider if you use it carelessly. Anthropic ran studies on this. depending on how you use the model, de-skilling in coding is measurable and real. The tool doesn't cause it, carelessness does.
6. The semiconductor space is his pick for a capitalistic win in the next decade. Physical world, traditional engineering, direct AI tailwind, not software but chips.
@aipost 🏴
⚡️Is Tesla secretly turning its charging stations into AI data centers?
Tesla quietly filed an application for MEGAPOD, a modular system designed for AI computing. On the surface, it looks like another piece of infrastructure. But the bigger idea could be far more ambitious.
Imagine Supercharger sites doing double duty: charging EVs during peak hours and supplying compute power for AI workloads when demand is low.
Tesla already has thousands of locations with power connections, networking, cooling expertise, and energy storage. Add AI hardware, and those sites start looking a lot like a distributed network of mini data centers.
If that vision becomes reality, Tesla wouldn’t just own a fleet of vehicles. It would own one of the world’s largest decentralized AI infrastructure networks.
@aipost 🏴
Yann LeCun says Elon Musk’s AI ambitions are already falling behind.
The Meta chief AI scientist called xAI a failure and predicted Musk’s AI push will ultimately collapse, arguing that the company can no longer keep pace with the industry’s leading labs.
LeCun pointed to reports that xAI is leasing capacity from its massive Colossus data centers to rivals such as Anthropic and Google, calling it a sign of weakness rather than strength.
He also argued that attracting top AI talent will become increasingly difficult, claiming Musk’s treatment of some of xAI’s original cofounders has hurt the company’s reputation among researchers.
@aipost 🏴
The AI prompt cheatsheet nobody taught you:
Role → Tell AI who to be
Context → Give background
Task → Be painfully specific
Format → List, table, or paragraph
Constraint → What to avoid
This structure cuts rework by 60-70%. Five lines. Better than 90% of prompts.
@aipost 🏴
The world's first Museum of AI Art has officially opened in Los Angeles.
DATALAND is now showcasing AI-generated work in a dedicated space. What started as a niche experiment has moved into a full museum setting, signaling a major shift toward mainstream acceptance.
The AI art era is already here.
@aipost 🏴
A significant cyber breach reportedly took place on June 11. On this day, sources indicate that an entity identified as “Mythos” managed to access almost all classified systems of the NSA and U.S. Cyber Command within a matter of hours.
Mark Warner, vice-chair of the Senate Intelligence Committee, stated that General Joshua Rudd, head of the NSA and Pentagon’s Cyber Command, confirmed the incident. According to Warner, the attacker broke into the classified systems at a much faster rate than typically seen in similar cases.
The Economist was cited as the origin of these statements. No additional details on the breach or its implications were provided.
📰 @aipost
🤖 OpenAI may have discovered a shortcut to making AI safer.
Researchers found that training models on realistic human situations didn’t just improve behavior in those specific scenarios, it made them behave better across completely different tasks.
The surprise? A model trained only on health-related interactions became more resistant to blackmail, deception, and reward-hacking in areas it had never seen before.
Even more interesting, OpenAI removed health and science data from training, yet the model still performed better on health evaluations. That suggests it wasn’t memorizing rules, it was learning broader habits like checking facts before making claims, admitting mistakes, resisting manipulation, and avoiding clever shortcuts.
The result was an AI that became harder to push toward harmful behavior while still staying helpful when given legitimate instructions.
In other words, OpenAI may be finding a way to teach models principles, not just rules.
@aipost 🏴
🤖 OpenAI just hired one of the architects of modern AI.
Noam Shazeer, a legendary AI researcher and one of the key minds behind the transformer revolution, has reportedly joined OpenAI to lead AI architecture research.
If his name sounds unfamiliar, his work definitely isn’t. Shazeer co-authored the 2017 paper “Attention Is All You Need”, which introduced transformers, the breakthrough that powers ChatGPT, Gemini, Claude, and nearly every major AI model today.
His résumé is stacked: multi-head attention, Mixture-of-Experts (MoE), and T5 all trace back to ideas he helped create.
After more than two decades at Google, Shazeer left in 2021 to found Character.AI. Google later paid a staggering $2.7 billion to bring him and the startup back, installing him as a key leader behind Gemini.
Now, one of Google’s most influential AI minds is heading to OpenAI.
@aipost 🏴
🦟 China built a drone the size of a mosquito
Researchers at the National University of Defense Technology unveiled a tiny drone that’s about 2 cm long and weighs less than 0.3 grams. Despite its size, it can carry a micro camera, microphone, and sensors that collect images, sound, and electronic signals.
The wild part? Its wings flap like a real insect instead of using propellers, making it look and sound much more like an actual mosquito.
Right now, it’s mostly a lab prototype. It can only fly for a few minutes, struggles in wind, and carries limited equipment.
But that’s the scary part. This is probably the worst this technology will ever be.
As batteries, sensors, and AI keep improving, insect-sized drones could eventually become powerful tools for surveillance, reconnaissance, and search missions.
@aipost 🏴
From Elegance to Sudden Death
@aipost 🏴
🇮🇳 The Indian workers training AI robots to take their jobs
With a smartphone strapped to her head, Indian housewife Nagireddy Sriramyachandra films herself performing everyday chores to train AI-powered humanoid robots to take on household jobs in the future.
@aipost 🏴
❗️Anthropic gets sued over its $200/month AI plan
A customer has filed a proposed class-action lawsuit against Anthropic, claiming the company’s premium Claude subscriptions didn’t deliver the level of usage that buyers expected.
The dispute centers on Anthropic’s Max 5x ($100/month) and Max 20x ($200/month) plans, which were marketed as offering much higher usage limits than the standard Pro tier. The lawsuit argues that users hit restrictions far sooner than the names and marketing suggested.
The case taps into a bigger AI industry problem: what exactly does “more usage” mean?
Unlike Netflix, AI subscriptions come with hidden constraints such as token budgets, rate limits, and compute availability. Two users paying the same price can get dramatically different amounts of work done depending on how they use the model.
Source.
@aipost 🏴
🗣Microsoft President Brad Smith says the recent backlash against AI from college students should be “a powerful wake-up call for the tech sector.”
Smith said young people aren’t against the technology itself, they’re worried about being replaced by it.
“To those in the tech sector who seemingly want to pursue a future where computers replace jobs, and AI becomes more capable than people, the next generation has offered a compelling response: ‘Not so fast,’”
He also said graduates are “telling us what we need to hear,” arguing that AI should help people do their jobs better, not take those jobs away.
@aipost 🏴