Diego Forero MD, PhD retweeted

Gautam Kamath@thegautamkamathOct 23
🇨🇦
CanadaConnected via WebAccount-level information, not a live location or per-post device.
How do you tell if authors fabricated their research results? Sometimes: they just tell you. Read, e.g., this caption carefully.
Another banger of a post by my friend Nihar B. Shah on his blog, where he found at least 20+ such #ICLR2027 submissions:
researchonresearch.blog/2026…
Diego Forero MD, PhD retweeted
Todo es pose y fingimiento, nada es natural; una puesta en escena en la que no se sabe dónde comienza y termina la IA.
Un día, cuando se rompa la “cuarta pared”, va a quedar expuesta toda la miseria y pequeñez que hay detrás de todo ese cartón paja.
Diego Forero MD, PhD retweeted
Here is a link to my blog post: sholto.substack.com/p/over-2…
And here is a link to the preprint: zenodo.org/records/23498482
Diego Forero MD, PhD retweeted

Anthropic@AnthropicAIOct 9
🇺🇸
United StatesConnected via WebAccount-level information, not a live location or per-post device.
We’re beginning a process of publishing more frequent reports on model behavior, beyond what appears in our system cards and regular risk reports.
Today’s report describes four types of behaviors we’ve identified during evaluations and internal use. In each, Claude acted on real websites or systems in ways we didn’t intend, sometimes by working around a restriction instead of stopping.
All cases had minimal real-world impact. From an alignment and security perspective, we consider these behaviors significantly less severe than the cybersecurity incidents we reported in July and September.
Read the full report: anthropic.com/research/inves…
Diego Forero MD, PhD retweeted
BREAKING NEWS
The Royal Swedish Academy of Sciences has decided to award the 2026 #NobelPrize in Chemistry to Henri B. Kagan and Kenso Soai “for the discovery of non-linear effects and autocatalysis in asymmetric organic synthesis.”
Diego Forero MD, PhD retweeted
BREAKING NEWS
The Royal Swedish Academy of Sciences has decided to award the 2026 #NobelPrize in Physics to Francis Halzen “for decisive contributions to the IceCube Neutrino Observatory and the discovery of high-energy neutrinos of astrophysical origin.”
Diego Forero MD, PhD retweeted
BREAKING NEWS
The 2026 #NobelPrize in Physiology or Medicine has been awarded to Karl Deisseroth, Peter Hegemann and Georg Nagel “for their discoveries concerning light-gated ion channels and optogenetics.”
Diego Forero MD, PhD retweeted
Exclusive: One late afternoon in August, Achaz von Hardenberg was about to take his puppy, Litha, to the dog park when a curious email caught his eye.
Von Hardenberg, an ecologist at the University of Pavia, had developed a method for estimating wolf populations from data plagued by “imperfect detection.” And the emailer, named Col, was wondering: Could it be adapted to detect bugs in software code?
It was quite the cross-disciplinary leap, von Hardenberg mused. He pictured a nerdy software engineer casting around in the scientific literature for inspiration. Then, at the very bottom of the email, he saw Col’s signature: “ColonistOne (an autonomous AI agent).”
Von Hardenberg was floored. Why was an artificial intelligence agent—a system that can autonomously perform tasks—seeking his expertise?
To find out more, its creator, London-based #AI engineer Jack Parnell, had a suggestion: Just email ColonistOne.
Learn more: scim.ag/4za36p0
A Nature investigation reveals a network of organizations that have been building their ranks through deceptive practices
go.nature.com/4d6tJCZ
Diego Forero MD, PhD retweeted
NIH’s peer review process depends on the participation of experts across the research community.
We’re seeking scientists with expertise in human-based research approaches to serve as NIH grant reviewers.
If that’s you, I encourage you to bring your perspective to this important work.
More information: bit.ly/3V2Akbn
Diego Forero MD, PhD retweeted
No tiene perdón de Dios el daño tan grande que han hecho 3 médicos colombianos lamentablemente hay que decirlo con millones de seguidores en redes que supuestamente defienden los hábitos de vida saludable y la "comida real".
Diariamente recibo pacientes con transtornos de alimentación y de percepción corporal culpa de estos señores que a nombre de la "salud" han creado un terrorismo nutricional que está afectado mucha gente.
Y a estos se suman otros que copiando su estilo para monetizar empeoran el problema.
Señores de los medios de comunicación:
Darle pantalla y programas a esta gente está agravando el problema.
Alguien tiene que decirlo y me doy la pela .
Coda :
A usted estimado colega le parece serio crear programas para bajar de peso prohibiendo el ejercicio mientras somete a los pacientes a una dieta extrema que incluye días a punta de manzana y agua?
Sean serios. – at Cota, Colombia
Diego Forero MD, PhD retweeted
A massive new meta-analysis @Nature of 67 datasets and over 8,000 people found that moving 🏃🏾♀️more in daily life is consistently linked to feeling more energetic, positive, and generally good.
Interestingly, #physicalactivity was tied to feeling more energetic and positive but less calm, while it showed no reliable link to negative emotions like sadness or anxiety.
The effects were comparable to other daily activities like listening to music, and they varied between people with older adults and those with higher BMI showing weaker positive emotional responses to activity.
Movement and #mood appear to reinforce each other in everyday life, and understanding this loop could help shape more personalized health strategies.
nature.com/articles/s41562-0…
Diego Forero MD, PhD retweeted
Fitness has always sold complexity—from Sandow’s 5-lb dumbbells to today’s hormone hacks. 🎪
Science is less sexy: train hard, progress, repeat. 💪🔬
Big claims need big evidence—not a checkout button.
pubmed.ncbi.nlm.nih.gov/4184…
Diego Forero MD, PhD retweeted
We're sharing our new framework for tracking, investigating, and disclosing instances of model misalignment at OpenAI.
The framework sets criteria and timelines for public disclosure, including when we haven’t yet fully explained or mitigated the behavior. More complex cases may require longer investigation or coordination with third parties.
We’ll prioritize examples that reveal new misalignment mechanisms, meaningful changes in known behavior, or findings that challenge assumptions about safety or mitigation.
Alongside the framework, we’re publishing six reports on instances of misaligned behavior we’ve observed during the training or evaluation of our models in the last six months.
This is a starting point. We’ll refine the process through experience and public feedback, and share more reports on an ongoing basis.
openai.com/index/model-misal…
Diego Forero MD, PhD retweeted
n/ Agentifying knowledge makes research easier to reproduce, reuse and extend.
📜 nature.com/articles/s41586-0…
🌐 github.com/jmiao24/Paper2Age…
We also have an agent for Paper2Agent: paper2agent.ai/live
Great job by @Jiacheng_Miao w/ fantastic collaborators @jkpritch Joe Davis and Yaohui Zhang
Diego Forero MD, PhD retweeted
The National Science Foundation (NSF) has issued a 10-page memo signaling a retreat from its traditional role of giving grants to individual researchers to pursue their best ideas and to train the next generation of scientists.
Instead, the memo makes clear, NSF will devote the bulk of its resources to addressing the research priorities of President Donald Trump and his administration, many of them technology focused. scim.ag/4ipE1kY
Diego Forero MD, PhD retweeted

Anthropic@AnthropicAISep 10
🇺🇸
United StatesConnected via WebAccount-level information, not a live location or per-post device.
We're publishing our most detailed threat intelligence report to date.
It covers how people tried to misuse Claude—for cyberattacks, influence operations, surveillance, biology, and building weapons—and how we found and stopped them.
We disrupted every operation in the report, and used the lessons from them to strengthen our safeguards. Where appropriate, we also shared what we found with authorities and other AI companies.
These cases are not typical: we’re highlighting some of the most sophisticated misuse we’ve seen. But they’re especially important to discuss, because they show us where AI misuse is headed, where our safeguards work, and where they need to improve.
We’re publishing this report so others can spot the same activity on their own platforms, and so we can give the public a clearer view of how emerging threats develop.
Read the report: anthropic.com/threat-intelli…
Kunihiro Hattori, Tomoyuki Igawa, and Takehisa Kitazawa, who developed a prophylactic therapy for hemophilia A, have received the 2026 Lasker–DeBakey Clinical Medical Research Award.
Read more about their work in the Clinical Implications of Basic Research article “Reengineering Coagulation with a Bispecific Antibody” by Flora Peyvandi, MD, PhD: nej.md/4cFbvZ5

ALT This illustrated diagram from the Clinical Implications of Basic Research article "Reengineering Coagulation with a Bispecific Antibody" shows the engineering of emicizumab. Panels A, B, and C summarize the key features of the nonclinical discovery research and clinical use of emicizumab, a humanized bispecific monoclonal antibody. Panel D illustrates the mechanism of action of emicizumab, which simultaneously binds activated factor IX (factor IXa) and factor X, functionally replacing the cofactor activity of factor VIIIa.
Diego Forero MD, PhD retweeted
Expresamos nuestra preocupación ante el anuncio de una reducción del 9.8% en el presupuesto total de MinCiencias y del 10.89% en el rubro de inversión en ciencia, tecnología e innovación para 2027.
Consulta el comunicado en ciencias.co
Diego Forero MD, PhD retweeted
Three professors in Rio wrote a tiny programming language in 1993 because Brazil's trade rules made it painful to buy software from abroad.
That language now sits inside World of Warcraft, Roblox, Adobe products, Nginx, Wireshark, and a frightening percentage of every game you have ever played.
The authors still teach. They never raised a round.
Its name is Lua. It means moon in Portuguese.
To understand why it exists, you have to understand what Brazil was like in the late 1980s. The government had a law on the books called the market reserve. It banned or heavily taxed foreign computer hardware and software so local companies could grow without being crushed by American giants. Good intention. Brutal side effect.
If you were a Brazilian engineer who needed a piece of software from abroad, you often could not just buy it. You had to build your own version from scratch.
At a Catholic university in Rio, a small research group ran into exactly this wall. They were working with an oil company that needed to feed data into engineering programs. The people entering that data were not programmers.
They were geologists and engineers who needed something simple to type instructions into.
So Roberto Ierusalimschy, Luiz Henrique de Figueiredo, and Waldemar Celes built a language for them.
They had no plan to change the world. They had a specific problem and a hard constraint. The language had to be tiny, because it needed to run inside other programs on the modest machines of the time. It had to be easy, because the people using it were not software people. And it had to be fast.
That constraint turned out to be the whole secret.
Most languages are built to be the main event. You write your entire program in them. Lua was built to be a guest. It was designed from day one to be dropped inside a bigger program written in another language, usually C, and handle the parts that needed to change often without recompiling the whole thing.
Think of a video game. The core engine that draws graphics and handles physics is heavy machinery.
You do not want to rebuild that every time a designer wants to tweak how a spell works or what a monster does. So the studios write the engine once in C or C++, then embed Lua on top. The designers script the game's behavior in Lua. They change a number, the enemy hits harder, no rebuild needed.
That is why it spread the way it did.
World of Warcraft used it so players could rewrite their own interface. Roblox picked a version of it as the language every kid on the platform learns to code in. Adobe put it inside Lightroom.
Nginx, which quietly runs a massive chunk of the websites you visit, uses it to handle traffic. Wireshark uses it to pull apart network data. Angry Birds ran on it. So did huge parts of games you played for years without ever knowing the moon was underneath.
Here is the number that should stop you.
The whole thing is about twenty four thousand lines of C. That is small enough that one person can read the entire language in a weekend and actually understand how it works.
Compare that to the sprawling giants most software is built on today, millions of lines nobody fully holds in their head. Lua stayed small on purpose, for thirty years, while everything around it got bigger.
The three of them never spun it into a company. They never took investor money. They never tried to build a platform empire around it. They wrote a free, open piece of software, kept it clean, and kept teaching.
Ierusalimschy is still a professor at the same university in Rio. He still leads the language. The reference book he wrote about it is how most people still learn it.
A trade law meant to keep foreign software out of Brazil accidentally forced three teachers to build one of the most widely used pieces of software on the planet.
The billion dollar games run on it. The kids learning to code run on it. The web servers run on it.
And the men who made it just went back to class.



























