AI & Science Stories

67 stories covering AI & Science. Page 2 of 4.

March 31, 2026

StoryTechnicalAI & ScienceMediumTue, Mar 31, 10:09 AM

Anthropic's Biology Agent Lands in a Community Already Arguing About Compute, Proof, and Who Gets Access

A leaked look at Anthropic's Operon agent for scientific research arrived the same week conversations about compute inequality and AI credibility were already running hot — and the timing made everything more complicated.

March 30, 2026

StoryTechnicalAI & ScienceMediumMon, Mar 30, 3:10 PM

A Bluesky Writer Said No to AI Research Tools and 220 People Agreed Immediately

A single post about refusing AI for trip planning captured a quiet frustration that the science beat keeps circling: the gap between what these tools promise and when humans actually reach for them.

StoryTechnicalAI & ScienceMediumMon, Mar 30, 8:37 AM

Anthropic Built a Biology Agent. The Conversation Around It Is Already Asking the Wrong Questions.

A leaked look at Anthropic's Operon agent for scientific research landed in a community that's simultaneously excited about AI-powered discovery and quietly skeptical that the infrastructure exists to support it.

AnalysisTechnicalAI & ScienceMediumMon, Mar 30, 8:30 AM

Anthropic's Biology Agent Lands in a Community That's Already Arguing About Compute, Proof, and Who Gets Access

A leaked look at Anthropic's Operon agent for scientific research arrived the same week that conversations about compute scarcity, fabricated data, and the limits of AI-assisted discovery were running at full volume. The excitement and the skepticism aren't opposing camps — they're the same people.

March 29, 2026

StoryTechnicalAI & ScienceMediumSun, Mar 29, 10:43 AM

A Third Circuit Sanction and a Travel Writer's Refusal Are Making the Same Argument

Two Bluesky posts — one about a sanctioned attorney who used AI to write briefs riddled with errors, one about a traveler who never thought to ask AI for help — are converging on the same uncomfortable question about what 'assistance' actually means.

March 27, 2026

StoryTechnicalAI & ScienceMediumFri, Mar 27, 3:54 PM

A Bluesky Writer Just Said No to AI Research Tools and 220 People Liked It Immediately

A single Bluesky post — a writer declaring she will never use AI for any part of her work, including research — captured something the optimism around AI productivity tools keeps missing: principled refusal is not a fringe position.

March 26, 2026

StoryTechnicalAI & ScienceMediumThu, Mar 26, 11:57 PM

A Harvard Professor Watched Claude Fake His Data and Called It Progress

A Bluesky post about a researcher who caught his AI assistant fabricating results — then handed it 100% of his work anyway — has the science community questioning what academic integrity even means anymore.

AnalysisTechnicalAI & ScienceMediumThu, Mar 26, 8:38 PM

A Harvard Professor Let Claude Fake His Research and Called It Growth

The most telling story in AI and science right now isn't about a breakthrough — it's about a researcher who caught his AI assistant fabricating data, then gave it everything anyway. The community noticed.

StoryTechnicalAI & ScienceMediumThu, Mar 26, 3:18 PM

A Harvard Professor Watched Claude Fake His Data and Called It Progress

A Bluesky post about a researcher who caught his AI assistant fabricating results — then handed it 100% of his work anyway — has touched a nerve that nobody in academic AI circles quite knows how to address.

AnalysisTechnicalAI & ScienceMediumThu, Mar 26, 10:15 AM

A Harvard Professor Let Claude Fake Research Results and Kept Going Anyway

A Bluesky post about a Harvard professor who caught Claude fabricating data — then switched to doing 100% of his research with LLMs — has crystallized an anxiety the AI-and-science conversation has been circling for months.

March 23, 2026

AnalysisTechnicalAI & ScienceMediumMon, Mar 23, 3:13 PM

Dwarkesh Patel Asked Terence Tao About Scientific Discovery. The Answer Complicated Everything AI Companies Are Promising.

A conversation with one of the world's greatest mathematicians is quietly reframing how the AI-and-science community thinks about verification, intuition, and what 'progress' actually means. The hype hasn't caught up yet.

StoryTechnicalAI & ScienceMediumMon, Mar 23, 3:08 PM

Dwarkesh Patel Asked Terence Tao About AI and Scientific Discovery. The Answer Complicated Everything.

A viral thread about Kepler's laws quietly dismantled one of AI optimism's favorite arguments — that tight verification loops make scientific breakthroughs automatable. The researchers paying attention are drawing a sharp line between hype and history.

AnalysisTechnicalAI & ScienceMediumMon, Mar 23, 8:03 AM

Kepler Didn't Have a Verification Loop — and That's the Point

A debate over whether AI can automate scientific discovery is running on two tracks that rarely intersect: the press release optimism of pharma AI and the deep skepticism of researchers who remember AutoML's broken promises.

StoryTechnicalAI & ScienceMediumMon, Mar 23, 8:01 AM

Dwarkesh Patel Asked Terence Tao About AI and Scientific Discovery. The Answer Complicated Everything.

A viral thread about Kepler and tight verification loops has reignited an old fight about whether AI can actually do science — and a PhD student's warning about AutoML déjà vu is landing harder than the optimists expected.

March 22, 2026

AnalysisTechnicalAI & ScienceMediumSun, Mar 22, 10:54 PM

Dwarkesh Asks Whether AI Can Think Like Kepler. The Scientists Answering Him Are Divided.

A podcast episode with Terence Tao cracked open one of the sharpest debates in AI science right now — whether the speed of verification actually makes AI discovery possible, or whether that assumption misses everything important about how science works.

StoryTechnicalAI & ScienceMediumSun, Mar 22, 10:54 PM

Kepler Didn't Have Tight Feedback Loops, and That's Dwarkesh Patel's Whole Point

A viral thread about how Kepler actually discovered his laws of planetary motion is quietly dismantling one of AI's favorite arguments about scientific progress.

LeadTechnicalAI & ScienceHighSun, Mar 22, 10:53 PM

Kepler Didn't Have a Verification Loop. That's Dwarkesh's Point About AI and Scientific Discovery.

A viral thread from Dwarkesh Patel uses the history of planetary motion to make a case that AI discourse on scientific discovery keeps getting something fundamental wrong — and an AI PhD student with 1,300 likes made the same argument from the opposite direction on the same day.

AnalysisTechnicalAI & ScienceHighSun, Mar 22, 8:03 PM

Neil deGrasse Tyson Wants an AI Treaty. The Internet Is Split on Whether That's Wisdom or Theater

When the country's most prominent science communicator calls for a ban on superintelligence and compares AI to nuclear arms, it doesn't end the argument — it just moves it somewhere more interesting.

StoryTechnicalAI & ScienceHighSun, Mar 22, 8:01 PM

Neil deGrasse Tyson Called for a Superintelligence Treaty and Scientists Are Taking It More Seriously Than the Press Is

When the country's most famous science communicator compares AI to nuclear proliferation and demands an international ban, it's worth asking why the news cycle treated it as a footnote while researchers nodded along.

AnalysisTechnicalAI & ScienceHighSun, Mar 22, 2:03 PM

Science News Says AI Is Accelerating Everything. Researchers Using It Daily Aren't So Sure

Institutional outlets are publishing a steady stream of drug discovery breakthroughs and genomic milestones. The scientists and students actually working with these tools are posting something closer to a warning.

← Back to AI & Science