[{"slug":"anthropic-quietly-built-a-biology-lab-to-test-its-ai-drug-ideas-in-real-life","title":"Anthropic Quietly Built a Biology Lab to Test Its AI Drug Ideas in Real Life","bodyHtml":"

Anthropic has quietly built a wet lab, a facility for physical experiments, in the San Francisco Bay Area, moving its biology work past computer simulations. The company wants to unlock treatments for rare diseases the pharmaceutical industry has largely ignored.

The timing is the strange part. Anthropic confirmed the lab just two weeks after its own researchers warned that AI could contribute to human extinction and reported that its systems could theoretically help with biological weapons development, warnings some scientists called overblown. The company is also gearing up for a roughly $2 trillion IPO. Life sciences head Eric Kauderer-Abrams insists the two goals aren't in tension: \"We're always balancing these things, making models better in ways that we think will uplift and accelerate developing medicines, and being thoughtful and careful about how we deploy it to limit the risk.\" A spokesperson later added, without elaborating, that the lab isn't for drug discovery specifically.

The company is chasing \"undruggable\" conditions using bispecific and trispecific antibodies, molecules that hit multiple targets at once, after buying startup Coefficient Bio for about $400 million in stock and adding Novartis CEO Vas Narasimhan to its board. Its August-launched Model Hardware Standard lets Claude direct lab robots directly, and Anthropic is now hiring procurement leads and protein specialists, while stopping short of running its own clinical trials so it doesn't compete with the pharma companies, Genentech, Bristol Myers Squibb, Novo Nordisk, that already buy its AI tools.

Why it matters: the same weeks Anthropic is warning the world that AI could be catastrophically dangerous, it's also building the physical infrastructure to make AI a hands-on part of inventing medicine, a bet that its safety concerns and its business ambitions can be managed side by side.

Source: Reuters

","excerpt":"Two weeks after warning that AI could contribute to human extinction, Anthropic quietly opened a real biology lab to put its AI to work inventing medicine.","seoTitle":"Anthropic Opens Biology Lab for AI Drug Discovery","metaDescription":"Anthropic has quietly built a physical biology lab in the Bay Area to test AI drug ideas in real experiments, just weeks after warning about AI's catastrophic risks.","source":"www.pexels.com","featuredImageUrl":"https://api.digitaldustlibrary.com/uploads/01a0ba20-94d0-7607-9f35-d6d90cd35c42.webp","categorySlug":"tech","categoryName":"Tech","categoryColor":"#C9553D","categoryFolderColor":null,"tags":[{"name":"AI","slug":"ai"},{"name":"Biotech","slug":"biotech"},{"name":"Health","slug":"health"}],"authorHandle":"hayk-baroyan","authorName":"Hayk Baroyan","publishedAt":"2026-09-19T14:47:22.216643+00:00","readingMinutes":1,"dispatchNumber":4,"format":"mini"},{"slug":"a-tiny-team-wants-to-send-a-spacecraft-to-alpha-centauri-it-will-arrive-in-80-000-years","title":"A Tiny Team Wants to Send a Spacecraft to Alpha Centauri It Will Arrive in 80,000 Years","bodyHtml":"

A small nonprofit company called the Fermi Explorer Mission wants to send a probe toward Alpha Centauri, the closest star system to the solar system, for less than $15 million. The team plans to launch within three years, funded almost entirely through public donations rather than a government space agency.

The catch is the timeline. At current propulsion technology, the spacecraft won't arrive for roughly 80,000 years long enough that whoever eventually receives its data, if anyone does, won't share a language with the people who sent it. The team behind it including Adi Oltean, co-founder of the orbital-data-center startup Starcloud actually expects a faster, later spacecraft to overtake it and arrive first. They're launching anyway. Someone has to go first, the reasoning goes, and a working mission, however slow, might push better-funded groups to try.

Donors get their name listed online for $10, or engraved on the spacecraft itself for $100. A full 81-page mission plan is public.

Why it matters: it's a real test of how far crowdfunding and a small team can stretch in a field usually reserved for national space agencies.

Source: TechSpot

","excerpt":"A small nonprofit wants to launch a probe toward Alpha Centauri within three years, funded almost entirely by public donations. It will arrive in about 80,000 years, and the team behind it already expects a faster, later spacecraft to beat it there. They're going anyway.","seoTitle":"This $15M Mission Aims for Alpha Centauri, 80,000 Years Away","metaDescription":"A crowdfunded nonprofit plans to launch a probe to Alpha Centauri within three years for under $15 million expected arrival: roughly 80,000 years.","source":"www.pexels.com","featuredImageUrl":"https://api.digitaldustlibrary.com/uploads/01a0615d-6eb4-7a4d-89a2-438e0f8f50f5.webp","categorySlug":"tech","categoryName":"Tech","categoryColor":"#C9553D","categoryFolderColor":null,"tags":[{"name":"Space","slug":"space"},{"name":"alpha centauri","slug":"alpha-centauri"},{"name":"exploration","slug":"exploration"},{"name":"Tech Culture","slug":"tech-culture"}],"authorHandle":"hayk-baroyan","authorName":"Hayk Baroyan","publishedAt":"2026-09-07T11:40:41.488406+00:00","readingMinutes":1,"dispatchNumber":3,"format":"mini"},{"slug":"chinese-banks-and-phone-carriers-are-now-giving-away-billions-of-ai-tokens","title":"Chinese Banks and Phone Carriers Are Now Giving Away Billions of AI Tokens","bodyHtml":"

Chinese banks and phone carriers have started treating AI tokens the raw units that pay for a chatbot's answers like a loyalty perk instead of a paid product. Moonshot AI launched a Kimi credit card with Agricultural Bank of China and American Express in July 2026 that pays cardholders in AI tokens and coding credits. China Merchants Bank offers up to 1.8 billion tokens for MiniMax's models to new cardholders, and Shanghai Pudong Development Bank goes further: up to 3 billion tokens for Alibaba's Qwen.

The giveaways work because tokens got absurdly cheap. Chinese models now cost 60-90% less to run than the equivalent from OpenAI or Anthropic, so handing out billions of them barely dents a bank's marketing budget. Daily token use across China jumped from 100 billion in early 2024 to roughly 500 trillion by mid-2026. One industry analyst called the whole push a \"supply-led experiment\" most people using these tokens still don't know what a token actually is.

Phone carriers followed the same logic. China Telecom sells plans from 9.9 yuan ($1.40) for 10 million tokens through its own TokenHub, which routes to 142 different AI models. China Mobile charges 1 yuan (14 cents) for 400,000 tokens, billed straight to a phone account. On the street, a Beijing bar throws in free DeepSeek access with a drink, and a dumpling restaurant hands out computing-credit vouchers after dinner.

Why it matters: in China, AI access is turning into a bundled freebie before most users even understand what they're being given a preview of how fast the cost of running AI could collapse everywhere else.

Source: https://restofworld.org

","excerpt":"In China, opening a new credit card or phone plan increasingly comes with free AI tokens thrown in a sign of just how cheap running AI has become there, whether or not customers know what a token is.","seoTitle":"China's Banks and Telcos Are Giving Away AI Tokens","metaDescription":"Chinese banks and phone carriers are rewarding customers with billions of free AI tokens, as local models get cheaper and daily usage rockets to 500 trillion.","source":"AI Modified","featuredImageUrl":"https://api.digitaldustlibrary.com/uploads/01a075a0-33bd-78a0-85a2-0d5d70476d94.webp","categorySlug":"tech","categoryName":"Tech","categoryColor":"#C9553D","categoryFolderColor":null,"tags":[{"name":"Tech Culture","slug":"tech-culture"},{"name":"World News","slug":"world-news"},{"name":"Business","slug":"business"}],"authorHandle":"hayk-baroyan","authorName":"Hayk Baroyan","publishedAt":"2026-09-07T11:40:26.512843+00:00","readingMinutes":1,"dispatchNumber":2,"format":"mini"},{"slug":"what-ai-has-actually-done-in-science-not-what-people-say-it-did","title":"What AI Has Actually Done in Science, Not What People Say It Did","bodyHtml":"

Forget the Argument, Check the Output

In September 2026, Anthropic announced that its Claude model had turned a famous, decades old mathematical proof into code a computer can check line by line. The proof was Andrew Wiles' 1995 solution to Fermat's Last Theorem, one of the most celebrated results in modern mathematics. Claude converted it into 13 million lines of Lean, a programming language built for formal verification, in eleven days. It did this by spinning up dozens of AI agents that chewed through 29,500 smaller theorems along the way [SiliconAngle]. Mathematicians had expected this kind of formalization to take years of very human, very caffeinated labor.

That's one data point. It doesn't prove AI is smart, useful, or safe, and it definitely doesn't settle the \"is this a bubble\" debate happening in every group chat right now. But here's the nice part: it's checkable. Anyone with the training can open the Lean file and confirm the code compiles and the proof holds, no faith required. That puts it in a different category from most AI claims you'll read today, which usually come from someone with skin in the game: a company selling the technology, a rival hoping it flops, or a critic who already wrote the \"I told you so\" post. This article sticks to what AI has actually produced in three fields of science, using only results other experts have checked, tested, or poked at with their own equipment. What you make of it is entirely up to you, not this article.

Why Science Is a Fair Test

Most arguments about AI happen in places where nobody has to prove anything: opinion columns, group chats, investor calls. Science plays by different rules. A mathematical proof compiles or it doesn't. A drug candidate kills bacteria in a lab dish or it doesn't. A cosmological model matches new telescope data or it doesn't. That makes science one of the rare arenas where an AI's output gets graded by people who don't care whether AI wins, only whether the specific claim is true.

The same pattern shows up in every example below: an AI proposes something, a proof, a molecule, a measurement, and human experts check it before anyone treats it as real. This isn't AI going off and discovering things alone in a locked room. It's closer to a very fast, very tireless intern who still needs a manager to sign off before anything ships.

Mathematics: Proof Checking and Problem Solving at Machine Speed

Formal mathematics happens to be a great match for AI, because a proof written in a language like Lean gets checked automatically by a computer. No arguing, no \"well, actually,\" just a compile that passes or fails. Claude's formalization of Fermat's Last Theorem is the clearest recent example. It didn't discover a new proof; it translated Wiles' existing one into a form a machine could verify, leaning on mathematician Kevin Buzzard's earlier groundwork and an open source planning tool called Prove2Me [SiliconAngle]. The result is now the largest Lean file ever written, which is either a proud achievement or a genuinely unreasonable file size, depending on your relationship with your code editor.

A separate, weirder result came from OpenAI, whose general purpose language model, not a system built specifically for math, disproved a 1946 conjecture by Paul Erdős about points on a flat surface spaced exactly one unit apart. The model linked the geometry problem to an unrelated field, algebraic number theory, and used a mathematical tool from 1964 to build a counterexample that simple grid arrangements never could reach. Nine mathematicians, including Tim Gowers, a Fields Medal winner (mathematics' version of an Olympic gold), checked and refined the work before it was published [Memeburn]. Disproving a conjecture, finding one case that breaks it, is generally easier than proving a new result true in every case, so call this impressive rather than earth shattering.

Both cases sit inside a wider trend. In 2024, Google DeepMind's AlphaProof and AlphaGeometry solved International Mathematical Olympiad problems at a silver medal standard, missing gold by a single point out of 609 competitors [DeepMind]. The systems needed each problem manually translated into formal language first, and took far longer than the four hours human teenagers get. Still, three separate labs producing checkable math results within about two years suggests an actual, repeatable skill, not a one off party trick.

$\"pexels-silverkblack-39190909.jpg\"$

Biology: From Predicting Structures to Proposing Drugs

Biology's clearest AI win predates this year by a while, but it's still the strongest card in the deck. In 2024, Demis Hassabis and John Jumper won half of the Nobel Prize in Chemistry for AlphaFold, an AI system that predicts a protein's three dimensional shape from its genetic sequence, a problem biologists had chipped away at for fifty years. (David Baker took the other half for solving the reverse problem: designing brand new proteins from scratch.) The AlphaFold database has grown from about 360,000 predicted structures in 2021 to over 200 million today, covering more than a million organisms, and over a million researchers have used it to study diseases including COVID 19 [EMBL].

A more recent case takes the same idea and points it at drug discovery. In April 2026, researchers at McMaster University, working with Stanford collaborators, used a generative AI model called SyntheMol RL to search through 46 billion possible chemical compounds, more than any human lab could ever physically screen, hunting for new antibiotics against drug resistant staph bacteria. The model proposed 79 candidates, and one, named synthecin, worked as a topical treatment against resistant staph infections in mice [Technology Networks]. The researchers are upfront that this is still an early step: nobody yet knows how synthecin actually works, it's only been tested on mice, and an approved medicine is still years away, so don't go looking for it at the pharmacy just yet.

Cosmology: Squeezing More Signal Out of Billion Dollar Data

pace telescopes and galaxy surveys cost a fortune to build, so squeezing more information out of data that's already been collected has real value. A team at Princeton University and the Flatiron Institute built a method called SimBIG, which trains AI models on millions of simulated universes and then turns them loose on real galaxy survey data. SimBIG found small scale galaxy clustering patterns that older statistical methods simply couldn't see, cutting the uncertainty in key cosmic measurements by more than half, roughly the same gain you'd get from analyzing four times as much data as was actually collected [The Brighter Side of News]. Given that a single galaxy survey can cost hundreds of millions of dollars, that's not a small trick.

A second team, from the Harvard Smithsonian Center for Astrophysics and the University of Tokyo, pointed a similar approach at dark energy, the mysterious force apparently speeding up the universe's expansion. They trained a neural network on light distortions caused by \"cannibal stars,\" stars that have snacked on a companion star, as seen by the Euclid Space Telescope, and used it to measure dark energy's effects about 10 percent more precisely than before [Archyde]. Neither team is claiming to have solved dark energy or ended cosmology's bigger arguments; both describe their AI methods as better tools for squeezing insight out of observations, with results that still need confirming from upcoming surveys.

$\"pexels-tomas-anunziata-129267-695477$

What These Cases Have in Common, and What They Don't Prove

Line up all six examples and the same shape keeps showing up. An AI proposes a proof, a molecule, or a measurement. A group of human experts, mathematicians, lab researchers, astronomers, then checks it before anyone treats it as real. None of these results happened because an AI worked alone in the dark; they happened because an AI did a mountain of preliminary legwork much faster than humans could, then handed the results to people qualified to say yes or no.

None of it is as tidy as the headlines make it sound, either. Formalizing a 1995 proof isn't the same as discovering new mathematics. Disproving a conjecture isn't the same as proving a hard new result. A compound that clears an infection in a mouse is not an approved drug, however hopeful the mouse looks. A 10 percent precision gain in one dark energy measurement doesn't settle cosmology's open arguments. Taken one at a time, each story is real but modest. Taken together, they show the same capability turning up across totally unrelated fields within about two years, and that part is worth paying attention to.

Decide From the Evidence, Not From Whoever Is Talking

Ask ten people whether AI is a bubble and you'll get eleven opinions, most delivered with total confidence. That's because almost everyone talking about AI has a horse in the race: the company selling it wants you amazed, the rival betting against it wants you embarrassed for being amazed, the pundit who called \"bubble\" years ago just wants to be right eventually. None of them are lying, exactly. They're just invested.

You're not being paid by any of them. Lucky you. So skip the shouting match and look at what nobody can argue with: a Lean file that compiles or doesn't, a mouse that gets better or doesn't, a measurement that holds up or doesn't. Six examples, six receipts, zero opinions required.

Be the annoying friend who checks the source at dinner table AI debates. Read what these systems actually did, try the tools yourself, and land wherever you land, bubble, breakthrough, or something boringly in between. It's your call, not anyone else's talking points.

References

Anthropic Uses Claude to Formalize Proof of Fermat's Last Theorem, SiliconAngle

AI Math Breakthrough: OpenAI Disproves 80-Year Erdős Conjecture, Memeburn

AI achieves silver-medal standard solving International Mathematical Olympiad problems, Google DeepMind

Computational protein design and protein structure prediction win Nobel Prize in Chemistry, EMBL

From Algorithms to Antimicrobials: AI Model Generates Novel Antibiotic Compound, Technology Networks

AI breakthrough unlocks hidden patterns in the universe's structure, The Brighter Side of News

Scientists Use AI & Cannibal Stars to Unlock Dark Energy's Cosmic Mystery, Archyde

","excerpt":"Forget the bubble debate. Here are six things AI has actually done in math, biology, and cosmology, all checked by real experts, none of them arguing with you about it.","seoTitle":"AI in Science: 6 Real Breakthroughs, Not Hype","metaDescription":"From formalizing Fermat's Last Theorem to designing new antibiotics, here are six verified ways AI has actually changed math, biology, and cosmology.","source":"www.pexels.com","featuredImageUrl":"https://api.digitaldustlibrary.com/uploads/01a076d4-4c56-79af-8e90-bac1559ef453.webp","categorySlug":"tech","categoryName":"Tech","categoryColor":"#C9553D","categoryFolderColor":null,"tags":[{"name":"Science","slug":"science"},{"name":"Mathematics","slug":"mathematics"},{"name":"Cosmology","slug":"cosmology"},{"name":"Biology","slug":"biology"},{"name":"AI","slug":"ai"}],"authorHandle":"hayk-baroyan","authorName":"Hayk Baroyan","publishedAt":"2026-09-07T11:40:16.627705+00:00","readingMinutes":8,"dispatchNumber":1,"format":"shelf"},{"slug":"tesla-s-cybercab-has-no-steering-wheel-total-recall-predicted-it-in-1990","title":"Tesla's Cybercab Has No Steering Wheel. Total Recall Predicted It in 1990.","bodyHtml":"

Tesla launched its Cybercab robotaxi in Austin on September 4, 2026: a car built from scratch with no steering wheel and no pedals, controlled entirely through a center touchscreen. It's now expanding to Dallas, Houston, Miami, Orlando, and Tampa.

When I was a young boy and saw Total Recall for the first time, there was a scene with a robot taxi driver named Johnnycab, played by Robert Picardo, who greeted passengers and drove them around without a human behind the wheel. I remember thinking it would be so cool to actually ride in a taxi like that one day. It hasn't reached Armenia yet, but watching Cybercab roll out in Austin, thirty-six years after that movie came out, feels like watching a piece of my childhood imagination turn into an actual product.

Cybercab runs on Tesla's Full Self-Driving software and seats two people on a single bench, with scissor doors that open upward and a trunk in the back but not the front. Every function, starting the trip, adjusting the seat, opening the doors, runs through the touchscreen; there's a cabin camera with a privacy light and an emergency stop button between the cabin lights. Elon Musk has floated a future price under $30,000, though Tesla hasn't confirmed a purchase price yet, only access through its Robotaxi app.

Why it matters: a robot taxi that felt like pure movie fantasy in 1990 is now a real product picking up real passengers in 2026, and if the pattern holds, plenty of other things that still feel like science fiction today are closer to arriving than we think.

","excerpt":"A childhood memory of Total Recall's robot taxi driver, and the strange feeling of watching Tesla's real one show up in Austin thirty-six years later.","seoTitle":"Tesla Cybercab vs. Total Recall's Johnny Cab, 1990–2026","metaDescription":"Tesla's driverless Cybercab launched in Austin with no steering wheel or pedals, decades after Total Recall imagined the same robot taxi on screen.","source":"AI Modified","featuredImageUrl":"https://api.digitaldustlibrary.com/uploads/01a07b9c-ac2c-7eb5-b048-23201b3270bb.webp","categorySlug":"tech","categoryName":"Tech","categoryColor":"#C9553D","categoryFolderColor":null,"tags":[{"name":"Culture","slug":"culture"},{"name":"Nostalgia","slug":"nostalgia"},{"name":"Tech","slug":"tech"}],"authorHandle":"hayk-baroyan","authorName":"Hayk Baroyan","publishedAt":"2026-09-07T11:24:41.798304+00:00","readingMinutes":1,"dispatchNumber":1,"format":"mini"}]

DIGITAL DUST Library

Tech

What AI Has Actually Done in Science, Not What People Say It Did