Jason Locasale @LocasaleLab

Nothing ages faster than an expert explaining why something will never work.

The peer review process has reviewed itself and found no problems.

The unfortunate irony is that Buttigieg served in an administration that encouraged the erosion of due process across our institutions. At the height of cancel culture, which his administration supported, accusations became sufficient to trigger investigations, public shaming, and professional ruin before any facts were established. If false accusations are unacceptable when directed at politicians, they should be unacceptable when directed at professors, students, scientists, employees, and ordinary citizens as well. Principles only matter if they apply to everyone. As anyone who has followed my account knows, I take false allegations that cause serious harm very seriously.

Pete Buttigieg @PeteButtigieg

11 hours ago

This week someone targeted my family for harm with a false report. We’re physically OK, but that doesn’t mean we weren’t harmed. I am beyond furious. Whatever your politics, this is awful, wrong, and can never become normal. substack.com/home/post/p-20…

6K 7K 43K 4.7M 1K

This is the longevity trial we need: Bryan Johnson vs. Freddy after three weeks of roadtripping and daily fast food. Measure everything.

Freddy🇩🇪 @FreddyLA7

8 hours ago

The drive up to Montreal and back to Vermont took up the whole day. Now we’re back and heading to the moose region in Vermont.🫎 Also, today is our 3-week anniversary since arriving in the US, and we’ve eaten fast food every single day. That has to be some kind of record.

1K 549 25K 530K 194

I like the framing of this take. I’ve said something similar in the past. Content should be evaluated based on the insight, originality, or entertainment it provides—not whether an LLM helped package it into a polished or familiar style. If the idea is genuinely interesting, who cares how the sentences were refined? Much of the objection feels stylistic, like insisting that real mathematicians should solve every problem by hand instead of using calculators. We eventually learned to value the quality of the thinking more than the mechanics of producing it.

vittorio @IterIntellectus

15 hours ago

my most controversial take is that it doesn't matter if a text is AI generated in the same way that an AI generated image can be art, AI generated code can be good code, math done on a calculator can be good math when it comes to texts, what matters is the idea expressed, and

89 31 647 26K 55

Basic science is more like the humanities than venture capital. Its greatest returns appear decades after the work was done. That is why it needs protection from bureaucracy and short-term thinking.

I completely agree. Six months ago or so, I feel could post something thoughtful that made people think and generated interesting discussion. Those posts would often reach a large audience. Now most of my posts seem to hit the same echo chamber and accumulate likes from the same people. If someone doesn’t consistently like my posts, future posts seem to disappear from their feed. Meanwhile, the posts that circulate widely are the edgy contrarian ones that generate outrage. It seems I need a large number of people calling me an idiot or a loser to get engagement. Something has to change. It’s making me seriously consider spending more time on other platforms.

Elon Musk @elonmusk

2 days ago

@karpathy @kuroke01 @gallabytes We need a complete overhaul of the algorithm

2K 889 13K 2.4M 787

This is exactly the kind of scientific credentialism that has done so much damage to science. Doudna won a Nobel Prize for work on bacterial CRISPR systems that enabled gene editing. That does not make her an authority on artificial intelligence, deep learning, or the future of computational biology. A Nobel Prize is not an all-access pass to pontificate on every scientific field. Notice what is missing from her argument: any evidence that she understands the current state of AI models, foundation models, reinforcement learning, or where the field is headed. Instead, we are told to accept her conclusion because she won a Nobel Prize. The claim that biology is too hard is particularly strange. That’s precisely why better computation matters. Nobody serious thinks today’s LLMs are independently curing disease. The question is whether increasingly capable AI, combined with better biological data, automated experimentation, and mechanistic insight from scientists, will transform biomedical discovery. The answer is almost certainly yes. The limiting factor is that biology remains one of the most data-starved and experimentally constrained sciences. As those constraints disappear, AI will become increasingly powerful. The biggest mistake here is assuming that because today’s systems haven’t solved biology, tomorrow’s won’t either. Scientists should know better than to extrapolate technological progress from a snapshot in time.

Ihtesham Ali @ihtesham2005

2 days ago

Jennifer Doudna won the Nobel Prize for gene editing and went on Bloomberg to say the chatbots everyone is betting on cannot innovate at all. Every promise Silicon Valley is making about AI curing disease just hit the one person qualified to check it. She has spent her whole

144 431 2K 293K 947

Paper mills are a symptom, not the disease. Paper mills are businesses that sell or facilitate scientific publications, often with minimal to no editorial or peer-review standards, to scientists seeking publications for career advancement. In a scientific system that truly rewarded truth, publishing in a paper mill would have no value. The market would disappear because nobody would care. Instead, our scientific institutions reward administrative metrics: publication counts, citation counts, h-indices, and journal prestige. Merit is reduced to numbers that are easy for bureaucracies to measure rather than scientific contributions that are difficult to evaluate. If careers depend on accumulating publications rather than advancing knowledge, someone will build a business that sells publications. Paper mills simply satisfy that demand. The solution is not more policing or compliance. It is to stop rewarding the metrics that created the market in the first place and return to evaluating scientists on the originality, rigor, reproducibility, and long-term impact of their work.

Retraction Watch @RetractionWatch

a day ago

“Paper Mills and the Fight Against Scientific Fraud.” insideprecisionmedicine.com/topics/precisi…

0 5 14 5K 2

Academic publishing is driven by corporate incentives: protecting subscription revenue, maximizing publication fees, maintaining readership, and preserving access to lucrative markets such as China. Journals behave like media companies competing for attention rather than institutions singularly focused on advancing truth. That encourages the publication of papers on fashionable topics like this one, politically useful narratives, and eye-catching claims that generate clicks rather than the most rigorous science. The frustrating part is that these same journals still publish some of the finest science in the world. Many of the most important discoveries appear in their pages. But editorial standards that reward attention over rigor diminish the credibility of the good work they publish.

Shivani Misra @ShivaniM_KC

2 days ago

What this demonstrates above all else is how broken peer review is

0 2 14 5K 2

One of the most important steps toward restoring merit is restoring our admiration for great scientists. Over the past decade, many of our most accomplished scientists were portrayed not as people to celebrate, but as elites, oppressors, or obstacles to be controlled. University administrations embraced this shift because it made it easier to consolidate power and elevate bureaucracy. Scientific progress depends on exceptional individuals with extraordinary creativity, insight, and persistence. We should celebrate them, honor them, and inspire the next generation to aspire to their level. As long as universities continue to marginalize their best scientists, talk of meaningful academic reform is impossible to take seriously.

Valeriy M., PhD, MBA, CQF @predict_addict

3 days ago

Godfather of LLMs Andrey Markov. spectrum.ieee.org/andrey-markov-… And now his book is available in English for the first time. valeman.gumroad.com/l/markov1900

0 13 80 6K 64

Nick articulates this important point well. This idea has even been mainstream in various forms in the health sciences, particularly in genetics, under concepts like precision medicine and N-of-1 medicine. Population averages—which is what you learn from a randomized controlled trial sampling a population—are not the same as understanding what will happen in a particular individual. The deeper our understanding of biological mechanisms becomes, the less we should expect average treatment effects derived from RCTs to be the final word. That is precisely why many of us advocate for more data, more assessments, and more personalized health. The goal is to move beyond relying on RCTs and population averages toward understanding the individual.

Nick Norwitz MD PhD @nicknorwitz

4 days ago

One of the biggest mistakes in medicine is assuming that a positive result in a peer-review RCT means a drug “will work for you.” It doesn’t. A statistically significant result tells you what happened, on average, in a population. It does not guarantee benefit for any given

23 11 111 14K 23

The best careers are built solving problems. The longest careers are built preserving them.

I think Laura’s major point is that psychiatric drugs are handed out like candy despite having life-altering consequences. All it often takes is a 15-minute meeting with someone basing their decision largely on self-reporting and guidelines from organizations with very little science behind them before a person is given drugs that can change the course of their life forever. I tend to believe that psychiatry is, in many respects, a soft science and that, at its worst, it can be weaponized for social control and the suppression of dissent.

Tyler Black, MD @tylerblack32

2 days ago

Laura Delano is straight up parroting Antipsychiatry here. The word psychiatric is unnecessary! Drugs can create physical dependence, which is why drugs need to be prescribed carefully.

7 6 54 7K 4

Chuck has been a big supporter of mine. I’m not sure what happened, but he’s a good guy. Hope this gets resolved quickly.

Walter Kirn @walterkirn

3 days ago

The suspension of @PezeshkiCharles is outrageous and recalls the worst days of COVID censorship on X. No cause has been given. @elonmusk

32 79 455 36K 15

If alcohol has no benefits, someone should explain how so many grant applications got written.

David Wolfe @DavidWolfe

3 days ago

Alcohol has no single health benefit. Not one.

2K 649 5K 1.6M 189

At Duke, where I used to work, tenure was sold during recruitment as a lifetime appointment with a guaranteed salary. The reality is that universities are far more capable of removing faculty they dislike than people realize. Administrators have an almost unlimited number of ways to marginalize and force out faculty they deem undesirable, especially productive faculty running large laboratories with greater compliance obligations and oversight. Resources disappear, responsibilities change, interference with grants is easy to accomplish and accumulates over time, and audits and investigations appear over routine matters. When these issues reach court, universities argue that tenure is merely a company policy rather than a legally binding commitment. Sally Kornbluth’s deposition, which is public, is worth reading on this subject.

Richard H. Ebright @R_H_Ebright

2 days ago

@LocasaleLab Tenure means the second step will rely on attrition and will require decades.

2 0 8 13K 1

Restoring scientists such as Carole Hooven who were pushed out for dissenting views, minor infractions, or simply failing to fit prevailing cultural and DEI schemes is an obvious and necessary first step in reforming universities. The next step is accountability. Universities must be held to merit-based standards and evidence. That would lead to a quick hollowing out of much of the faculty they have recruited, promoted, and retained over the past ten years.

National Association of Scholars @NASorg

5 days ago

Three prominent scholars are calling on Harvard to re-hire evolutionary biologist Carole Hooven, calling her cancellation over defending biological sex an act of intellectual vandalism. Restoring her would send a powerful message against cancel culture. thecollegefix.com/professors-ask…

18 403 2K 58K 77

The future arrives as a bad idea according to the people in charge of the present.

@vijaypande Interesting. The curvature probably says something about reproducibility/generalizability too. A large upward or negative value would indicate the paper is taking off or losing interest.

My point is that these metrics like h-index are misleading. Two obvious problems with the h-index come to mind first and there are many others. First, it favors older scientists, especially those who have not done much recently and have spent years or decades letting citations accumulate. Second, it does not account for whether papers are actually correct. A paper can be wrong and highly cited. A paper can send an entire field in the wrong direction for years and still be highly cited. Both contribute to a high h-index.

Wafik S. El-Deiry, MD, PhD, FACP @weldeiry

4 days ago

@LocasaleLab Congrats. People criticize it but true peers appreciate it. And it doesn’t stand alone.

1 0 9 8K 0