SCROLL NEWS / DISCOVERY
Search headlines
Find the stories shaping the conversation.
Results for " Only 3 " 32 found 🔀 AI-powered Shuffle
Venture Capital Investments in Maternal Health Startups
A new study by LDI Fellows identified 172 VC-backed maternal health startups, but fewer than half accepted insurance and only 34% took Medicaid.
LEGO Batman: Legacy of the Dark Knight is 30fps on Nintendo Switch 2
Some Nintendo Switch 2 owners have been disappointed to find out that LEGO Batman: Legacy of the Dark Knight is only 30fps on the Nintendo Switch 2. The latest LEGO game doesn’t feature a Performance…
Americans Care About Their Oral Health, But Few Cite Whole Body Health as Associated. Crest and Oral-B Are Here to Change That.
New Survey from P&G and American Academy of Family Physicians Reveals 76% of Americans Say They Care About Their Oral Health, But When Polled, Only 3% of Americans Associate Oral Health with Whole Body Health
Americans' satisfaction with schools is at a low point, new Gallup poll shows
A new poll shows that only 32% of Americans are satisfied with the quality of K-12 education in the country | 1080 WTIC NEWSTALK
Americans' satisfaction with schools is at a low point, new Gallup poll shows
A new poll shows that only 32% of Americans are satisfied with the quality of K-12 education in the country.
Who Won August and September: Original Films or Franchises?
(Welcome to the Entertainment Strategy Guy, a newsletter on the entertainment industry and business strategy. I write a weekly Streaming Ratings Report and a bi-weekly strategy column, along with occasional deep dives into other topics, like today’s article. Please subscribe.) I just heard someone say that if you’re reading a pundit or data analyst, and they don’t tell you that they got anything wrong, they’re not a pundit but an “influencer”. I agree with that. You have to tell your audience when you get things wrong, or they won’t (or shouldn’t) trust you as much. Today, we’ve got another edition of “What I Got Right, What I Got Wrong”. In general, I’m patting myself on the back (and patting very hard) about a few things like IP at the box office, LIV golf, horror films, crowdfunding movies, superheroes, and old people going to the theaters. That said, I also made a couple of data mistakes on the streaming bubble popping and HBO’s datecdotes, so I’m not perfect. But first, I need your feedback... Subscribe Follow-Ups: What Should We Call Mid-Budget Movies Which Aren’t Cinematic Enough for Theaters But Are Too Expensive to Make Money on Streaming? After I asked for feedback on what we should call straight-to-streaming films that are too big to pencil out on streaming but not really big enough to resonate in theaters, I got some great suggestions from you all, including... “Moldilocks” from John Aboud “Little Big Indie” from Travis Frick1 “Midflicks” from Jonathan Funke. “Extra-Medium” from Jona Nwuke. (Read the explanation in the footnote.) Between these and Brandon Katz—who suggested “Bermuda Budget Triangle”, “The Platform Gap” and “The Distribution Deadzone”—we’ve got some excellent suggestions. And they’re better than my suggestion, the “Streaming Budget Dead Zone”, so let’s take a poll! The winning entry becomes the new term. Loading... RIGHT: LIV Golf is Fully Bankrupt…Is Anyone Else Next? A few years ago, I was always a bit perplexed at all the articles I’d read praising LIV Golf and their strategy to disrupt the PGA. LIV Golf, a brand new professional golf league, offered PGA stars ten times what they made on the PGA Tour to join their new league. And folks praised this strategy for its initial “success” in that a lot of big names did indeed leave the PGA. Yeah, of course the players came over. LIV Golf paid them so, so, so much more! But LIV Golf hadn’t discovered a way to increase potential revenue. So think about this in basic business terms… They paid much more in costs… …but had no real way to increase revenue. That’s not a strategy! That’s deficit financing. It won’t work unless you bankrupt the competition (and then turn around and pay those same golfers much, much less). It’s not a sustainable strategy and only lasts as long as the company/person/nation backing it decides they’re cool with losing money. For LIV Golf, that meant Middle East oil wealth, in particular Saudi Arabian money. Clearly, the current Iran War has hurt Middle East finances, so they need to trim their more exorbitant spending. And LIV Golf was part of that trimming. This should be a major warning for others. TGL’s parent company, TMRW Sports, just got a $1 billion valuation, despite TGL’s (the indoor golf league) horrible TV viewership of less than half a million viewers per match. Unrivaled, the women’s basketball league, just secured a $650 valuation, despite its horrible TV viewership and a new rival (Project B) entering the scene next year. To relate this to streaming, Hollywood should ask which streamers may have wealthy patrons funding their losses. (Let’s be clear: Google, Amazon and Apple.) Could those patrons lose their appetite? For Amazon, probably not. But Apple has a new boss, so maybe! WRONG: My Analysis of the Streaming Bubble Was Missing a Week I’m going to be congratulating myself a lot today, but I make mistakes too, like this data goof. When I last compiled the data on the decline in TV shows, I was missing one week at the end of June. If you look at the first image, you can see that one week was mislabelled as “July”. Mainly, this image got updated to show a 27% decline, not 29%: This change doesn’t really impact the overall analysis, but it is two percent better. By the way, through the third quarter, the decline increased and it’s now a 30% decrease since 2022. (I’ll write/visualize this in an upcoming article.) But I want you to trust me and my data. Especially these days, when many people are using LLMs that I know are inserting faulty data into their charts, I want to keep earning my audience’s faith. So that’s the accurate data. RIGHT: IP Remains Very, Very Popular In July, I wrote a giant (and I mean giant) article on Backrooms, Obsession, and the box office, going over what we know, what we don’t know about what works in theaters, looking at YouTubers, IP, the horror genre, comic book movies, and a whole lot more. The month of August really tested a lot of my theses and, being honest, mostly supported my arguments. Let’s start with IP. Looking at 8-Aug (the weekend after Spider-Man: Brand New Day came out) to 18-Sep (the weekend that Resident Evil came out, which I think provides a nice bookend to this time period), there were seven films based on pre-existing IP: Resident Evil (2026) ($126 million) Practical Magic 2 ($65 million) _Insidious: Out of the Furthe_r ($65 million) Coyote Vs. Acme ($59 million) Paw Patrol: The Dino Movie ($53 million) Tony ($15 million) Super Troopers 3 ($7 million) Compare those to the notable original films from the past month—I actually could have included more movies, but here are just thirteen, bringing us to an even twenty films—including.... The End of Oak Street ($54 million) Buddy ($26 million) Mutiny ($15 million) By Any Means ($15 million) The Dog Stars ($14 million) Runner ($14 million) One Night Only ($11 million) Hope ($8 million) Spa Weekend ($7 million) The Uprising ($6 million) Teenage Sex and Death at Camp Miasma ($6 million) Onslaught ($3 million) Eli Roth’s Ice Cream Man ($2.8 million) Here’s that in chart form: Five of the top six films in this time period were all based on IP. I made a big chart of films that grossed over $200 million at the box office before Spider-Man: Brand New Day and The Odyssey hit theaters. Let’s update that chart! By the way, if you want to see how I categorized each film—so another bar chart—here it is: I know that many of my fellow critics/pundits/analysts dislike films based on IP and how Hollywood is making so many of them. And I’m sympathetic to this point of view. As I’ve written many, many times before, you need a balance between existing franchises, new IP, and original films. And Hollywood clearly needs to make more films like Resident Evil (a well-made film from a visionary director) and fewer Practical Magic 2’s (which didn’t get critical or customer buzz). But at some point, critics and pundits need to contend with what audiences are telling them: Movie-goers aren’t showing up to original films. Audiences are speaking with their dollars, telling you they want more IP and franchises. You can try to convince studio heads to make fewer IP-based films and franchises, but the data and numbers aren’t there. Instead, critics need to work harder to convince audiences to show up for original films. Aim your ire/concern at the average person, not studio heads.2 Because they’re just making the films that audiences are telling them to make. WRONG: Original Horror Films Didn’t Break Out I’ll be honest, even though I wrote an article casting some skepticism on the horror genre in July, if you asked me to make a prediction, I would have predicted that, in August, a new, original horror film would have blown up. No, seriously, I just assumed that Obsession and Backrooms presaged a change in audience behavior. But none of the buzzy new original horror films from August—Teenage Sex and Death at Camp Miasma, Onslaught (not an action film in spite of the ads), The End of Oak Street, Eli Roth’s Ice Cream Man, or _Buddy—_broke out. To be clear, exactly one of those films (Buddy) had good “ROI”, but again—I try to be specific in my language—none were “popular” in any broad sense of the word. None of them will be “saving” movies theaters like Backrooms or _Obsession_helped save the summer. The new Insidious film and Resident Evil, both based on IP, were far and away the biggest horror films since July. (We’ll see if this changes in October/Halloween season.) RIGHT: Stay Skeptical about Crowdfunding... I’ve long been skeptical about crowd-investing platforms as one of Hollywood’s saviors, mainly because there’s so much hype/buzz. People need to stay more skeptical about more things, explaining both the potential upside but also the downsides. In particular, the media often hypes crowdfunding at the start and never checks in on the actual results after they’ve come in later. And August gave us our first update! _Ice Cream Man—_directed by Eli Roth—grossed $6 million off of a $5.5 million budget. This is a production of The Horror Section, which was one of the first “crowd investing” studios with 2,400 investors, which means that 2,400 investors probably lost money. They certainly aren’t getting as great of returns as if they had just invested their dollars in the stock market. Hopefully Stiletto (Tagline: “Someone’s Going to Make it Rain Blood!”) does better next month. WRONG: Another Data Goof Here’s another data error. When the first episode of House of the Dragon came out, HBO put out that it had 21.5 million viewers in the first three days, and I read that to mean in the US…but no, it was global. So my US-only datecdotes charts shouldn’t have included it. We never got US-only numbers for the first episode, but for the final episode, HBO put out that it had 11 million US viewers. (And that global dropped to 21 million viewers.) Here’s the updated chart (which I’ve since used in the Streaming Ratings Report): Still, this show is absolutely huge. RIGHT: Superhero Films Remain Very, Very Popular After Supergirl flopped, I read a few takes that “comic book movies are going the way of the Western”. Post-Spider-Man: Brand New Day, that take didn’t age well. To be fair, I have a very nuanced take on the superhero genre right now; it’s down right now, for a lot of reasons. But it’s not “dead”. Maybe _Spider-_Man is just a really popular character? I saw that take, and it’s a fair counter-argument. (But pundits arguing that superhero movies were dead should have mentioned this $1.5 billion counter-argument…) But is it just Spider-Man? The next Avengers film already has $50 million in pre-sales (and I was skeptical that that film would do well) and the Avengers: End Game re-release topped the box office two weekends ago (over three original films). And I wouldn’t bet against Batman or Superman. So maybe it’s just Spider-Man, Batman, Superman and the Avengers. Oh, and Deadpool, of course. And Black Panther. And Wolverine. And probably the X-Men. Plus a well-made Wonder Woman or the Hulk film could break out. But that’s it! It’s just those ten characters/teams. Oh, what’s that? Lanterns is also doing well on HBO? (See previous section…) To be fair, I’m actually pretty sympathetic to the argument that more popular characters—like Spider-Man and Batman—anchor more popular films. In fact, I made that exact argument three years ago when I first wrote about the “Marvel-cession”. In many ways, you can blame The Guardians of the Galaxy for fooling Marvel Studios (and the rest of us) into believing that any character could pop. It turns out, the list of iconic characters is probably smaller than most people think. But it’s probably too early to say that superhero films and comic book movies are dead unless “death” means a slight decline over a longtime. RIGHT: Who Killed Theaters? Old People I get frustrated whenever I see headlines or analysis about how young people are “returning” to theaters. As I’ve detailed(for years), young people have always powered the US box office, despite narratives about “kids these days” and their “phones”. Really, what’s changed post-2020/pandemic is that old people aren’t going to the movies nearly as much. This summer, I saw a movie (from an older director) in a theater near a retirement community, and multiple older people at the theater were talking about how this was their first time seeing a movie in years. I dislike personal anecdotes, so YouGov can fill in the data, best summarized by this headline: “Who killed movie theaters? Not the youths”. According to them, 64% of people aged 18-29 have seen a movie in the last year, but only 30% of 65-and-older. 20% of 18-29 have seen a movie in theaters in the last week and 42% in the last month. Here’s the polling data: Most concerning? Many Americans (17%) think theaters are a worse or much worse experience than watching films at home. Slight WRONG: Hadestown Opens Big A live theater capture of the Broadway musical, Hadestown, made $20 million at the US box office, which begs the question: was I wrong to be skeptical about musicals a few years ago? Yes and no. On the one hand, $20 million is a far cry from being “popular”, so yeah, the genre isn’t that popular overall and Hadestown is one of the more popular musicals from recent years (i.e. the “Taylor Swift Data Fallacy” in action). On the other, I doubt filming this cost all that much, and I don’t think that they spent much on marketing, so this is a good source of ancillary revenue. Smaller Updates WRONG: As I mentioned in a Streaming Ratings Report, I underestimated the budget for Enola Holmes 3. It probably cost more like $50 million, if not more. But... I’m not sure that it really matters? At sub-10 million hours, prices have to come down to make this work. WRONG: Netflix is giving Ink a 27-day in theaters! To quote the kids/YouTubers these days, let’s go! Now I might actually have a chance to see Danny Boyle’s latest in theaters. I’d complained about this in a “Coming Soon” section, but I was heartened to read that Netflix is giving multiple films longer theatrical windows this year. RIGHT: Netflix is sending 4-5 films per year to theaters. Netflix is slowly but surely sending more and more films to theaters, as I cautiously predicted earlier this year. For now, it’s just three big films and a number of awards contenders, but still, this is great news. And they’ll be releasing box office grosses! Just this week, Ted Sarandos confirmed that KPop Demon Hunters 2 will come to theaters (and my guess is it performs in the box office top ten at a minimum). WRONG: Angel has 3 million subscribers! How do I know this? Well, they told Deadline, who reported it. I marked this as “wrong”, since they’ve doubled their subscribers in one year but, you know, they don’t really have a hit film to speak of and they’re still losing money. WRONG: Furious was only renewed for one more season. I accidentally wrote “two more seasons” in my latest “Renewals, Cancellations, Un-Orders and Removals Update”. RIGHT: House of David is ending with its third season. In July, Prime Video renewed House of David for a third season, as I just wrote in my latest “Renewals and Cancellations” report. Well, now it’s ending after that third season. Why? As I’ve been writing, its viewership wasn’t great. I got feedback that this show didn’t cost very much, but it cost enough that its limited viewership didn’t save it. WRONG: Adults was a Hulu original! So I missed Adults when it first came out last year; I saw that it aired on FX and just assumed that it was a linear-first program. Turns out, it aired three episodes on FX, but binge-released the rest of its episodes the next day on Hulu. Huh. So I should have covered it last year! But I just wrote about it. 1 “When I was a kid in the 90s, if you wore a t-shirt to school that wasn’t too big and also wasn’t too small, but somehow didn’t quite fit, we’d say you were wearing an “extra-medium” shirt.” 2 As always, a huge exception is Disney, which barely makes anything original anymore.
PCOS postpones perimenopause and allows pregnancies at older ages
Only 3 per cent of those with polycystic ovary syndrome reach perimenopause by the age of 46, which may allow them to conceive when older
Americans give Trump low marks on the economy in new poll : The NPR Politics Podcast
Only 36% of Americans approve of President Trump’s handling of the economy in a new NPR/PBS News/Marist poll. We discuss what might be driving that discontent and how much Trump’s tariff policies are to blame. This episode: senior White House correspondent Tamara Keith, White House correspondent Danielle Kurtzleben, and senior political editor and correspondent Domenico Montanaro.This podcast was produced by Casey Morell and Bria Suggs, and edited by Rachel Baye.Our executive producer is Muthoni Muturi.Listen to every episode of the NPR Politics Podcast sponsor-free, unlock access to bonus episodes with more from the NPR Politics team, and support public media when you sign up for The NPR Politics Podcast+ at plus.npr.org/politics.
Microsoft calls Copilot ‘entertainment only’ while charging $30 a month for it
Microsoft's Copilot Terms of Use label it "for entertainment purposes only", yet the company charges up to $30/user/month and has spent $80bn on AI. Only 3.3% of users are paying.
Reedijk 'not an academic guy' despite doctorate
New Hibernian manager Marink Reedijk insists he is more a man of action than an academic despite having a philosophy doctorate in sports coaching. The Dutchman is only 35 but had already worked in coaching with Ajax and West Bromwich Albion, Roeselare, Vitesse and Anderlecht before leading Beveren to the Belgian second-tier title and promotion last season. "I'm not a really an academic guy," he told Hibs TV.
The Science Behind the AI Panic Is Shakier Than You Think
A column by Sascha Lobo Is AI making us dumb? So asked DER SPIEGEL on its cover a while ago. After a few weeks of recovery and a therapeutic intervention with the help of my AI, I feel able to respond. That’s a bit unfair, of course; the article itself is more nuanced than the loud cover, the image of a schoolkid with dull prompt questions, or the subtitle: “AI is conquering the schools, and educators fear a disaster.” The short answer to “Is AI making us dumb?” is: no. The long answer is more interesting, and even carries a preliminary glimmer of the spectacular. Unfortunately, it requires a detour into the sphere of artificial intelligence, into the mogul-field terrain of technology assessment, the attention economy, and the destructive virus of cultural pessimism. Along with narrowing the scope to AI’s effects on education and work rather than on intelligence. Because the honest version of the question would be: Does AI help us learn and work, or not? Generative AI, which set off the still-ongoing hype with ChatGPT, has only existed in relevant distribution since late 2022, early 2023. For a technology-driven transformation of people and society, that is extremely short. For that reason alone, research is still in its infancy. And technology assessment is not a fast discipline anyway. Prominent voices from Elizabeth Eisenstein to RAND to Tom Wheeler say we still cannot conclusively judge the consequences of the printing press. Its invention is a mere 580 years back, and some scholars would rather not circulate assessments that might have to be retracted in two or three hundred years. We’re Going Through Growing Pains A phenomenon that already caused upheaval in social media research makes serious assessment of AI’s consequences even harder: the “moving target” problem of LLM-ology, as the linguist Sean Trott called it in 2025. AI develops so quickly and so unpredictably that researchers often end up studying snapshots of the past. There are concepts, from morphological analysis to anticipatory governance, for keeping research, planning, and regulation halfway meaningful anyway. But even with those, universally valid statements are rare. DER SPIEGEL - The German View is a reader-supported publication. To receive new posts and support our work, consider becoming a free or paid subscriber. Subscribe We should therefore treat findings about the interplay between humans and AI as far more procedural and situational. Unfortunately, this challenge comes at the worst possible time. We are going through growing pains: AI’s everyday impact has grown very large very fast, and with it the longing for guidance. Annoyingly, the body of solid knowledge has not exploded along with it. Experts cannot be moved to broad consensus on even banal statements about AI. Sascha Lobo, born in 1975, is an author and strategy consultant focusing on the internet and digital technologies. Together with Jule Lobo, he explores the debates of the day in the podcast “Feel the News – Was Deutschland bewegt” (”What Moves Germany”). Nvidia CEO Jensen Huang says software developers are dying out. Bill Gates considers coders one of only three professions that will survive the AI tsunami. Top tech economists find virtually irrefutable arguments both for and against an AI bubble. The same naturally applies to the structurally youth-hostile debate over whether AI is good or bad for young people. The gap between the demand for insight and the supply of solid knowledge creates an orientation vacuum. This vacuum of AI cluelessness is filled, as the attention economy described by Georg Franck in 1998 would have it, with half-knowledge (funnily enough, often AI-generated), marketing, and self-marketing. After all, infinite money is at stake, along with academia’s best current shot at global fame. But this iridescent froth of interpretation operates more by the algorithmic principles of the digital public sphere than by seriously earned insight. That is why dramatized, fear-feeding claims about AI are so promising. It tempts parts of academia into distorted drivel, and parts of the media then sensationalize the published drivel once more. Was the Research Aimed at a Desired Result? How badly this can influence academia and public perception is demonstrated by one of the most-discussed studies on AI’s effects: “Your Brain on ChatGPT.” The MIT study by Nataliya Kosmyna also gets plenty of space in the SPIEGEL cover story. What was measured and interpreted: lower brain activity in people who write with AI assistance, compared to people who only google or write without any tools. As is common with preprints, the study has not yet been peer-reviewed, and it contains substantial weaknesses that experts have sharply criticized. The first round had 54 participants in three groups, the second only 18, who reportedly also came from MIT’s own orbit. There was no preregistration (committing in advance to what you want to find out with which data). In their analysis, Stanković et al. even discovered an opposite effect that the study does not further explain. Some measurement data has not (yet) been published. And the EEG analysis method used is considered too imprecise to support such far-reaching conclusions. The biggest criticism, though, is the suspicion of an activism-driven study. In an interview with Time, lead author Kosmyna said she published the study immediately, rather than after full peer review, because she feared politicians might otherwise decide to put ChatGPT in kindergartens. She considers that absolutely harmful. It is legitimate to hold such a view. It just seems unscientific to approach a study with it, and then to design the setup so that the chance of a fitting desired result remains as high as possible. Ideologies Collide It is not easy to separate genuine AI insights from studies that trivialize and weaken well-founded AI criticism. The same goes for the opposing, AI-enthusiastic narratives, because the pro-AI side works against it with billions in resources. Ideologies collide, fed by egos, PR strategies, moral posturing, and once-in-a-lifetime opportunities. Doomsday marketing makes the search even harder: AI companies can profit from warnings about the dangers of AI, whether through the notoriety of apocalypse narratives or because, in some target groups, danger translates into power. Some seemingly negative findings are therefore quite welcome in the AI industry. If you try to get at the findings presumably less influenced by interest groups, the picture in education seems relatively clear: AI can improve educational outcomes when the right AI is used the right way, namely pedagogically. Simply dumping an AI chatbot into a classroom, by contrast, accomplishes little. Huge surprise. But something else is also obvious: whether they are allowed to or not, kids use AI for school, which is why the controlled integration of AI tools into education is, in the medium term, entirely without alternative. In the world of work, the landscape of findings is more complex and requires more differentiation. A series of serious studies points to an AI effect that suits neither the alarmists, regulation fetishists, and AI opponents, nor the trillion-dollar corporations with their armada of paid scientists, PR professionals, and ecosystems: working with AI helps the inexperienced (at least for now) considerably more than it helps the pros, in certain fields of work, and provided they know how to use it correctly. Missing Domain Knowledge Among the experienced, the professionals, and the high flyers, by contrast, despite frequent time savings, partially negative effects and new problems are emerging. Take the jagged frontier problem: AI solves task A brilliantly and fails completely at task B, even though both look similar even to professionals. Or the verification-cost problem, in which AI’s greater productivity gets eaten up by the enormous effort of checking its output. Or the illusion of competence, because AI makes new domains seem easily penetrable, until you painfully discover that deep domain knowledge is missing. Taken as a whole, the effects of working with AI do lean positive. But the promise that AI works as a turbocharger for everyone and boosts practically everything simply refuses to materialize. The AI magic the world has been feeling since ChatGPT arrived in late 2022 has so far barely shown up in most companies’ numbers. The consulting firm McKinsey has coined a term for this, the “gen AI paradox,” and believes that work processes and corporate structures would have to be rebuilt quite fundamentally to realize AI’s benefits at scale. Reuters just reported that only 3.3 percent of Microsoft’s customers have subscribed to the AI package Copilot. That may be down to the product; in other areas we continue to see soaring successes and new AI triumphs. But after the first quarter of 2026, disillusionment is the word for all too many AI projects at traditional companies: because the promised revolution of AI agents has so far failed to arrive or been postponed, and because of the leveling effect that many of the studies cited here suggest. Against Inequality The societal effects, on the other hand, appear rather encouraging. Many previous technologies operated on the Matthew principle, named for the Gospel verse about giving to those who already have: they tended to reward the more competent and more privileged. AI, however, seems more often to have an equalizing effect once a threshold of access and education is crossed. Some anecdotal insights point in this direction, for example, how an unemployed man in Leipzig successfully defended himself in court in 2025 against fraud accusations from Germany’s federal employment agency — without a lawyer, armed only with ChatGPT. The AI made mistakes and cited invented court rulings, but in the end the man walked away unpunished. And a key capability of artificial intelligence is its capacity to learn: set up correctly and fed with data, it can keep getting better. It is a question of time, or more precisely of AI innovations and training data, until even free AI chatbots write better legal briefs than law graduates with perfect exam scores. Whether that will make anyone dumb is something the world’s press will surely inform us of in due time. Subscribe Leave a comment
AI needs science’s search history
“Actually, that’s where the gold is,” said Alasdair Russell, my graduate school friend who leads a pre-clinical genome editing group at Cambridge. He was talking about the winding road of science that is omitted from published papers. “When you’re discussing how to do the experiment, and why this way is better than another way, and what does the data really mean? I know what it shows, but what does it mean?” At RAAIS 2026, he described how his group has begun logging the twisting path of discovery as it happens, recording the verbal and written exchanges that normally disappear. Ideas become nodes in a living graph: they branch when a meeting produces two plausible experiments, merge when separate lines of evidence converge, and go dark when someone quietly stops pursuing them. In his implementation, one agent scores novelty, while another tries to learn “how scientists think and how they navigate through a complex world of data,” so that high-potential nodes can trigger deeper investigation. As frontier AI labs deploy agents toward scientific discovery, giving AI authentic scientific taste remains a trillion-dollar dilemma. The bottleneck is that the record we have kept for centuries of science might be insufficient to get us there. The experiments that never make the paper Peter Medawar, 1960 Nobel laureate and “father of transplantation,” asked in 1963 whether the scientific paper was a fraud, and answered yes: the form of the paper misrepresents the thinking that produced it. Sixty years on, the diagnosis is unchanged. A paper presents a clean progression from hypothesis to result to conclusion. Lost along the way are the unconventional theories, the abandoned or unaffordable methods, and the underwhelming and inconclusive data. It reads like a browser history with every dead end deleted: the ten open tabs, the four rephrased queries, the wrong turn down a subforum. All scrubbed, leaving one clean path from question to answer as the canonical path. What has changed since then is that the discarded material is now worth something. No high-impact journal wants these artifacts, but raw trial and error is the most likely source of the training data required to develop scientific intuition, or what researchers call taste. Earlier attempts to capture the discards were motivated by scientific integrity. The Journal of Negative Results in Biomedicine launched in 2002 to publish rigorous studies that disputed established models or exposed ineffective treatments. Its archive preserved negative conclusions after they had become papers, rather than the live alternatives and arguments that produced them. BioMed Central closed it in 2017, saying the mission had been served now that other journals publish null results. A less generous reading is that in fifteen years it published around 200 papers because almost nobody wants to read a negative result, let alone write one up when it will not count toward academic tenure. AI models are different. Even an uninteresting negative result can be useful, provided it is labeled. Earlier this year, Anthropic put a version of this to the test. The company pulled 129 decision points from real Claude Code sessions between January and March 2026, showed models the work up to a human detour, and asked what should happen next. Claude Mythos Preview beat the human choice 64% of the time, while Opus 4.5 managed 51%. The comparison was tilted toward the models because Anthropic deliberately chose moments where the human decision had room for improvement. On 127 further scenarios where the human action was already strong, the models improved on it only about 20% of the time. The study was possible because Anthropic’s researchers work inside a tool that logs by default. The reasoning, detours and outcome are produced in the same working environment. Biology has no equivalent. Its reasoning happens in hallways, at benches, in Slack threads and on whiteboards, while the outcome arrives weeks later somewhere else. The record has to emerge from the work itself. Ask scientists to reconstruct it afterwards and we’ll create another polished account. Subscribe A log is not a label Before any of this becomes training data, a negative result has to say what failed. The first paper published in the Journal of Negative Results in Biomedicine examined 234 negative studies across five leading medical journals. Only 30% discussed statistical power, and half clearly defined a primary outcome. Tell a model an experiment failed, without telling it whether the assay was underpowered, a reagent had degraded, or the hypothesis was simply wrong, and it will learn noise with confidence. Decision histories have a second missing label too. When a lab considers five experiments and runs one, only the chosen branch returns an outcome. The other four are experimental counterfactuals. Researchers call this the selective labels problem: the data reveal results only for the actions someone chose to take. Such a record can teach a model to imitate a lab’s taste, but it cannot establish that the taste was good. It also smuggles local constraints into general lessons. A lab that never ran cryo-EM because it did not own a microscope teaches a model that cryo-EM is rarely the right call. Alasdair’s nodes preserve the candidate set, which is the necessary first step. To be useful, each decision record needs six fields: The evidence available at that moment, and nothing that arrived later. The candidates considered, including the ones dismissed in a sentence. For each candidate: the expected result, the confidence attached to it, and the cost in time and money. The route chosen, and the reason. What happened next. How the evidence changed the scientist’s view. These must be timestamped before the outcome, because hindsight turns uncertainty into inevitability. Disagreement also has to survive. Averaging three scientists into one clean rationale reproduces exactly the information loss we are trying to fix. But there is an obvious failure mode here. Once decisions are logged and scored, people log for the record post-facto. Anyone who has watched an electronic lab notebook fill up with retrospective tidying knows how quickly a research tool becomes a compliance exercise. A record that costs a scientist ten minutes of honest reflection per decision is worth far more than one that costs an hour of performance. A publication initiative called Registered Reports offers a useful starting point. Here, researchers submit their rationale, methods and analysis plan for peer review before the data exist, and the journal commits in principle to publish if they follow the approved plan. Nature has now expanded the format across every field it covers. This approach to paper writing timestamps intent before the outcome, but it freezes one plan. A useful search history must also preserve how the plan changed, which alternatives were rejected, when, and on what evidence. Subscribe Run the runner-up Autonomous labs show what happens when decisions and outcomes are connected in a loop. Liverpool’s mobile robotic chemist ran 688 experiments over eight days in a ten-variable formulation space. A batched Bayesian search used each result to choose the next experiments and found photocatalyst mixtures six times more active than the starting formulations. Every completed experiment changed what the system did next. The robot optimized within a goal and search space that humans had already chosen. It did not decide which scientific question mattered. So what this work demonstrates is narrower, but still useful: a search history pays off when the options return standardized outcomes quickly. Open-ended biology is harder because branches can take weeks and the discarded alternatives may never be run. Some exploration therefore has to be bought. Where two branches are plausible and the stakes justify the cost, a lab should sometimes run the runner-up. Otherwise the record captures what today’s scientists usually chose and stays silent on what they systematically overlooked. A funder could ring-fence a small fraction of a grant for the branch not taken, provided that the decision record and the outcome are both deposited. That costs money, but so does having every lab rediscover the same abandoned path. Test whether taste transfers When it comes time to evaluate if our new AI system exhibits taste, the comparison should run across laboratories and fields. At selected decision points in a live research campaign, we should freeze the six fields: the evidence, the candidates, expected results, confidence, costs and the choice made. Experts then rank the options before the outcomes are known, and those outcomes are checked independently later. We could give models one of two diets - papers alone, or papers plus decision histories - and ask them to rank the next experiment on unfamiliar projects. Then, we score information gained per dollar and week, calibration, and how quickly weak branches are abandoned. If histories improve choices only inside the lab that produced them, they amount to useful organizational memory. If they improve choices on unfamiliar problems elsewhere, Alasdair’s group will have captured something science has never managed to write down: a transferable record of taste. Until that happens, a scientific search history is a promising record, not yet a training set.
Cincinnati Reds' Brady Singer Gets Honest About Recent Struggles
The right-hander was charged with the loss on Thursday after giving up nine hits and seven earned runs in only 3.1 innings pitched.
Why Fenerbahce turned to Europe’s youngest sporting director to break their title drought
Devin Ozek is only 30 but, having played his part in Bayer Leverkusen's rise, swapped Germany for the pressure-cooker of Istanbul football
Quetta earn first win as poor Kingsmen lose again
Quetta Gladiators earned their first win of the Pakistan Super League season by inflicting a second defeat on Hyderabad Kingsmen. Kingsmen, beaten by 69 runs by Lahore Qalandars in their opening game, lost by 40 runs in Lahore as they made a poor attempt at chasing 175. After taking 17 from the first over, Marnus Labuschagne's side managed only 34 from the next eight with only one boundary coming between the end of the second over and the start of the 10th.
Americans have elected only 3 Black governors. This year, 8 major party candidates are on the ballot
With Florida Republicans nominating U.S. Rep. Byron Donalds on Tuesday, there are eight Black major party gubernatorial nominees — men and women, Democrats and Republicans — on November ballots.
The N2 Company Earns Inc.'s Best in Business Distinction for Social Good
The N2 Company is proud to announce its inclusion on the Inc. Best in Business list in the Best Social Good category. Inc.'s annual Best in Business Awards recognized only 32 companies for Social Good — companies that go beyond financial success to make a meaningful difference through their products, initiatives, and impact-driven leadership.
Beach parking, money in politics and more: Four San Diego measures move one step closer to November ballot
Support was unanimous for three of the measures, but the committee vote was only 3-2 in favor of a measure to ban paid parking at beaches and bays.
Americans' satisfaction with schools is at a low point, new Gallup poll shows
A new poll shows that only 32% of Americans are satisfied with the quality of K-12 education in the country | KNSS Radio
Women and girls in science: Dismantling barriers, closing gender gaps
Although women are more likely than young men to pursue higher education, they make up only 35 per cent of science graduates.
Premier League: Is top-flight possession football on way out?
Only 36% of teams with more than 60% possession have won this season in the Premier League. BBC Sport looks at why?
Lamar Alexander reflects on six decades in American Politics
As Tennessee prepares to elect a new governor in just 76 days, Alexander, who was only 38 years old when he first won the governorship in 1978, is reflecting on his massive marathon of public service.
Americans' satisfaction with schools is at a low point, new Gallup poll shows
A new poll shows that only 32% of Americans are satisfied with the quality of K-12 education in the country
Debt Digest | Federal Debt is a Health Care Problem
Welcome David Ditch to the Debt Dispatch! We are pleased to announce a new regular contributor: David Ditch will be joining the Cato Institute as a Policy Analyst for Budget and Entitlement Policy later this week. David has more than a decade of experience in fiscal policy issues, with a particular focus on federal spending. He previously worked at the Senate Budget Committee, the Heritage Foundation, and the Economic Policy Innovation Center. David’s areas of expertise include appropriations, transportation, agriculture, federalism, grantmaking, and policy options for deficit reduction. Originally from the Rochester, N.Y. area, David has a BA in Economics and Political Science from the University of Rochester and a MA in Political Management from George Washington University. David currently lives in Arlington, VA. You may have enjoyed his earlier guest posts: CBO: New Highway Bill Has More Spending, Taxes, and Deficits Trump Administration’s Proposed Changes to Federal Grants Highlight Problems of Big Government Here are this week’s reading links and fiscal facts: The federal debt is largely a health care problem. In a new report, the Congressional Budget Office and Joint Committee on Taxation find that “In 2026, federal subsidies for health insurance, net of related payments to the government, are projected to equal $2.4 trillion, or 7.4 percent of gross domestic product (GDP). In CBO and JCT’s projections, those subsidies grow by 65 percent, reaching $3.9 trillion, or 8.4 percent of GDP, in 2036. Subsidies for Medicare contribute most to that overall growth, increasing by $900 billion. Over the entire 2026–2036 period, federal subsidies for health insurance total $33.6 trillion.” Cato’s Michael Cannon emphasizes, “The long-term federal debt problem is a health care problem. […] Only two categories of federal outlays will grow faster than gross domestic product (GDP): health care subsidies and interest payments on the debt. The former is, therefore, the primary driver of the latter.” The bond market fears deficits, not the loss of tariff revenue. Responding to a New York Times essay claiming tariff revenue has grown too important for a future administration to unwind, Cato’s Kyle Handley argues: “Bond investors are not attached to customs duties as a line item revenue source. They care about the government’s overall fiscal position.” He continues, “tariff revenue is a side hustle. And the Trump administration has already promised to dole out the funds through schemes like tariff dividend rebates, farm subsidies, and pay-fors on tax cuts or other spending.” Furthermore, “tariff revenue is simply not large enough to transform the government’s fiscal trajectory.” Take net interest for example: it “reached $970 billion in fiscal year 2025, absorbing 18.5 percent of federal receipts. Customs duties accounted for only 3.7 percent of receipts—and that was before refunds (see Figure 2).” Handley concludes, “The real bond-market concern is a large and growing interest bill, persistent budget deficits, and a political system unwilling to bring spending and revenue into alignment.” Medicare’s spending growth is driven by more volume and utilization. A Congressional Research Service report finds that Medicare spending “grew at an average annual rate of 7.3%” from 1985 through 2025, and the trustees project health care expenditures will keep rising “faster than gross domestic product (GDP) in most future years.” Citing CBO, the report attributes Medicare’s projected 2026–2036 spending growth to “23% from higher enrollment, 31% from inflation, and 47% from the growth of inflation-adjusted spending per beneficiary. In other words**, the largest single driver of Medicare spending growth during the next decade is expected to be higher volume and intensity of health care services.”** Boccia and Thakur explain: “as the economy grows, Medicare spending tends to grow at least as fast—and often faster—because the program automatically pays the bill for more and more healthcare consumption by seniors, even as prices rise.” The Fed has little control over market interest rates. Cato’s Jai Kedia documents that the Fed has held its rate target steady since December, yet “nearly every rate Americans actually borrow at has climbed over the same stretch.” Kedia continues, “The Fed did nothing, and the cost of credit went up anyway. […] Markets spent these months repricing macroeconomic events such as sticky core inflation, a volatile energy market driven by the conflict in the Middle East, and global trade disrupted by tariffs, among others. None of that required a policy change to show up in borrowing costs because markets, not the FOMC, set prices.” He concludes, “Accurately priced borrowing rates will come from credible disinflation and disciplined budgets and from a Fed content to follow the economy rather than pretend it leads it.” Nearly half of US farm income comes from taxpayers. Former Reagan OMB director David Stockman notes that “in the most recently completed full year (FY 2025) farmer incomes in the US totaled $97.8 billion, but fully 42.7% or $41.8 billion of that amount came from taxpayers, not the marketplace. Nor is this some kind of one-year aberration. If we look at the most recent decade as a whole, total farm income posted at nearly $708 billion, but, as indicated, fully $322 billion or 45% of this came compliments of US taxpayers.” Cato’s Chris Edwards explains why so little of that income is market-earned: “Farmers are businesspeople, but the government shields them from just about every type of weather and market risk. Furthermore, just about every part of the agricultural industry is subsidized, including insurance, loans, marketing, research, export sales, and land improvements.” The Debt Dispatch is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Subscribe
Doing Business in ... Woodinville
Woodinville is a young city, incorporated only 33 years ago, but it has aged well — not unlike the wine for which it is widely known.
Only 3% of students meet standards in science: National assessment data revealed
Partial data from Israel’s 2025 Meitzav national assessment exams has been revealed, showing that only 3% of ninth-grade students met the benchmark in the science exam, while 54% failed.
Bad vs. worse in politics
If you ever thought the President Donald Trump-led Republican Party is bad, well the Democrats are worse. The Democrats lack politically astute leaders, as well as vision and discipline. They seem incapable of being on offense and are always on defense where it is difficult to score points and win. Only 37 percent of Americans […]
Texas Continues To Face Shortage Of Mental Health Professionals As Demand Rises
Texas faces a mental health workforce shortage in 2026, with 393 HPSAs affecting 13.4 million residents and only 32% of needs met.
'Bridal arms' are trending: This simple 3-move Pilates workout sculpts your shoulders, back and biceps
Here's a short routine with only 3 movements to try
Scientists catch a hidden electronic state forming in just 30 femtoseconds
Scientists watched a light-triggered hidden state form inside a material in only 30 femtoseconds, revealing a step that had never been seen before. The material first entered a fleeting electronic state in which its bonds reorganized in a repeating pattern, followed by tiny atomic shifts. This ultrafast pathway could offer a new way to control electronic properties with light and help inspire faster, more responsive technologies.
Meet the man whose groundbreaking book is being devoured by the left
Anton Jäger has been lauded for his exceptional ability to make sense of these strange political times – and he’s only 32. The writer and academic explains why we’re living in an age of hyperpolitics
Texas A&M Offense Faces Major Test Against LSU’s Dominant Defense
Texas A&M must protect Marcel Reed, establish the run and finish drives against an LSU defense allowing only 33 rushing yards per game.