💸🤖 Most paid research positions in AI quietly require years of prior experience. Anthropic just opened one that does not.
The Anthropic Fellows Program pays $3,850 a week for 4 months of full-time research, plus around $15,000 a month in compute funding and direct mentorship from company researchers.
Candidates need to be fluent in Python and have a strong technical background, but no prior AI research is required.
Fellows are paired with mentors and aim to produce a public output such as a paper. In previous cohorts, 25 to 50% of fellows received a full-time offer at the company.
A few years ago, getting close to frontier AI research required a PhD and years inside a major lab. Now Anthropic is paying people without that background to spend 4 months trying.
Full details and application here: job-boards.greenhouse.io/anthropic/jobs/5023394008
What are your thoughts on this? 🤔💬 —
➡️ If you want to keep up with all the AI news, useful tips, and important developments, join 100k+ subscribers reading our free newsletter
🤖👀 A graphic designer just reminded everyone what creating digital art used to look like before AI tools arrived.
Adham Sameh (@TheAdhamSameh ), an Egyptian designer known for the elaborate composite artworks he creates for Bleacher Report Football, shared a throwback to a process built on dozens of carefully selected reference photos.
No AI tools, no stock libraries, no shortcuts. Each figure was built piece by piece, with every fold, shadow, and texture matched by hand in Photoshop.
AI image tools now compress that workflow into a single prompt and a few seconds of generation.
The precision and authorship behind manual composite art looks very different from what these systems produce.
His post captures a shift the design world is still adjusting to.
What are your thoughts on this? 🤔💬
🤖🎯 Pigeons kept landing on his balcony, so he built a computer vision turret to deal with them.
The setup pairs a USB camera with a small $100 mini computer. When the camera sees a pigeon, an AI model running on the board tells two small motors where to aim a battery powered water gun, and a simple circuit pulls the trigger.
The clever part is the AI model. It is what is called open vocabulary, so the target is not locked in. You type what you want it to spot and the system tracks that. Pigeons today, anything else tomorrow.
The full build is a mini computer, a camera, two tiny motors, a few electrical components, and a toy water gun.
Everything runs on the device itself with no cloud and no internet required.
A few years ago this would have needed expensive hardware and custom training. Now it runs on a cheap board with a free model pulled off the internet.
What are your thoughts on this? 🤔💬
Source: Muxamilian on Reddit
—
➡️ If you want to keep up with all the AI news, useful tips, and important developments, join 100k+ subscribers reading our free newsletter
🤖💰 NVIDIA's market cap just passed $5.7 trillion, putting the chipmaker above the entire GDP of every country on Earth except the United States and China.
The milestone was reached on May 14, 2026, after the stock surged more than 4% to a new intraday record.
For context, Germany, the world's third largest economy, has a nominal GDP near $5.45 trillion in 2026. Japan sits close to $4.46 trillion. NVIDIA is now larger than both.
Comparing market cap to GDP is not a direct comparison. Market cap is the total value of NVIDIA's shares at a single moment, based on future earnings expectations. GDP measures the actual output a country produces in one year.
Different units, but the contrast still shows how much value markets now assign to a single company relative to entire economies.
The driver is the same as it has been for two years.
Demand for AI chips keeps outrunning supply, and hyperscalers continue to commit massive capital to data center buildouts. NVIDIA reported $215.9 billion in revenue in fiscal 2026, up 65% year over year.
NVIDIA produces this valuation with roughly 36,000 employees. Germany's economy runs on a workforce of around 46 million. The contrast shows how concentrated the value of the AI buildout has become inside a small number of companies.
What are your thoughts on this? 🤔💬
—
➡️ If you want to keep up with all the AI news, useful tips, and important developments, join 100k+ subscribers reading our free newsletter
#ai #artificialintelligence #technology
NVIDIA CEO Jensen Huang broke away from President Trump’s schedule and headed straight into one of Beijing’s traditional hutongs to enjoy a bowl of zhajiangmian (noodles with soybean paste) from the famous Fangzhuan workshop.
His spot on the trip to China was not guaranteed. Huang was initially left off the delegation and added at the last minute, joining executives like Tim Cook and Elon Musk. The timing matters because NVIDIA sits at the center of the U.S. and China chip standoff, with its market share in the country dropping from around 95% to nearly 0% under export restrictions.
A high-stakes summit, the CEO of the world's most valuable company, and a bowl of street noodles in the middle of it all.
What are your thoughts on this? 🤔💬
🧹👤 47 data broker listings gone. 12 dead accounts deleted. 3 search results buried. All in one weekend, all with one chatbot.
A user on X going by digital ghost shared how he cleaned years of personal data off the internet using Claude in about 6 hours.
The model drafted formal CCPA opt-out letters for brokers like Spokeo and BeenVerified, wrote GDPR and KVKK deletion requests with proper legal language, mapped the full removal process for old accounts found through justdeleteme.xyz, and ranked breaches / by risk to decide on password changes, 2FA, or full deletion.
For articles that could not be removed, the strategy shifted to outranking them with new content.
Most of this work used to live with lawyers and developers. Now it lives in a chat window.
What are your thoughts on this? 🤔💬
This man uses AI to put himself into Squid Game and replaces every elimination with a new scene.
Now the scariest thing is a loud "mwah" sound effect followed by a man in a hoodie grabbing your face. @advkiki put himself into every major scene and turned the deadliest game show on television into an involuntary group therapy session.
We've all watched shows wishing someone would just step in. AI is letting creators actually do that now, and the results blend with the original footage way better than they should.
The wildest part is there's no studio, no production team, no VFX house. It's just one creator with AI tools, a hoodie, a tie, and absolutely no respect for the tension Squid Game spent hours building. 🎥: @advkiki
What are your thoughts on this? 🤔💬
Well well well, a viral post on X just turned into one of the sharpest tests of how people actually judge AI art.
On May 12, the account SHL0MS uploaded a real Claude Monet water lily painting and told followers it had been generated by AI, asking users to explain what made it inferior to a genuine Monet.
Hundreds responded with confident critiques, including self-described art critics and verified accounts. One said the water looked like it was on fire. Another called the work emotionless. A third offered a technical breakdown of the tree reflections.
Then the account revealed the image was a real Monet, and a number of the critiques quietly disappeared.
Research published in Scientific Reports has shown the same pattern before, with people rating art lower when it is labeled as AI made, even when they cannot tell it apart from human work.
What are your thoughts on this? 🤔💬
🦸♂️🤖 A solo creator just built the Superman and Invincible crossover that licensed studios will never make, and it runs as a full cinematic AI film.
The project was made with Seedance 2.0 alongside original editing, sound design, and direction.
Baby Kal-El crashes on Viltrum and grows up under the rule of the conquering super race from Invincible, moving through volcanic wars, cosmic invasions, and mysterious forces from deep space as a weapon of empire rather than a symbol of hope.
A crossover like this used to require a full VFX studio and a serious budget. Now a single creator can build it with a video model and a clear vision.
The interesting AI filmmaking is no longer demos of what a model can render. It is creators telling the stories licensed studios will never make.
What are your thoughts on this? 🤔💬
Source: @multiverse.of.ai67
Someone on X reportedly got banned from Claude after trying to use it to help create a hantavirus vaccine. 👀🧬
The viral posts quickly sparked a much bigger debate around AI safety and biology.
Users tested Claude with prompts related to vaccines, pathogens, and virology research. Some reported chats being paused by safety systems, while others claimed their accounts were restricted after repeatedly pushing those boundaries.
Anthropic has strict safeguards around viruses, genetic engineering, lab workflows, and biological experimentation. And honestly, this highlights one of the biggest tensions in AI right now.
These models are becoming powerful enough to discuss advanced biology and medical research at a surprisingly high level. But the more capable they become, the harder it gets to separate harmless curiosity from potentially dangerous misuse.
What started as memes on X ended up showing how seriously AI labs are treating biological risk.
What are your thoughts on this? 🤔💬
🤖₿ A Bitcoin holder just recovered around 5 BTC, worth $400K, from a wallet he had been locked out of for 9 years, with help from Anthropic's Claude.
After 8 weeks of testing roughly 3.5 trillion passwords with btcrecover and Hashcat on rented GPUs, nothing worked. As a last attempt, he dumped his entire old college computer into Claude.
The model found an old 2019 wallet backup, matched it with a seed phrase from a college notebook, and used that older password to decrypt it.
Since Bitcoin private keys never change, those same keys still controlled the current wallet. Claude also debugged an error inside btcrecover where a shared key and password were being concatenated in the wrong order.
The 5 BTC, untouched since April 2015, were swept out the same day.
What are your thoughts on this? 🤔💬
—
➡️ If you want to keep up with all the AI news, useful tips, and important developments, join 100k+ subscribers reading our free newsletter
(Prompt below) Russian influencer Sasha Kriletskaya (@alexandrakriletskayaa ) has gone viral by joining the trend of the moment: using AI to create videos that look like real scenes from live broadcasts. But she took the idea to another level.
She wrote: "I hope I have ended the trend. ”
To create everything, she generated the images in ChatGPT Images and then animated the scenes using China's Seedance 2.0.
Prompt:
Ultra-realistic sports broadcast still of a glamorous woman sitting in a packed football stadium crowd during a night match, wearing a dark brown sleeveless high-neck satin top and black square earrings, shoulder-length light brown/blonde hair styled in soft waves. She is casually drinking from a tall blue aluminum can while holding a half-eaten cheeseburger in the other hand. Around her are fans in bright yellow and blue football jerseys and scarves, creating strong team-color contrast. The scene feels candid and cinematic, captured mid-game from a TV broadcast camera angle with shallow depth of field. Include realistic stadium seating, crowded audience atmosphere, broadcast overlay graphics in the top-left corner showing a live football score and match timer, and a sports network watermark in the top-right. Natural arena lighting, detailed skin texture, sharp focus on the woman, slightly blurred background crowd, authentic live sports broadcast aesthetic, 16:9 composition.
