Paul Kaschner retweeted
Ich bleibe dabei: Private Spendenaktionen, z. B. zum Geburtstag, sind eine großartige Möglichkeit, Organisationen wie @MECFSResearch oder @weandmecfs zu unterstützen und gleichzeitig mehr Aufmerksamkeit für ME/CFS zu schaffen.
Heute werde ich 23 und verbringe meinen Geburtstag wegen ME/CFS im Bett.
Viele von euch hier wissen leider selbst nur zu gut, was diese Erkrankung bedeutet.
Deshalb habe ich eine Spendenaktion für die Mecfs Research Foundation gestartet.
Und bereits jetzt sind über 1000€ innerhalb von einem Tag zusammengekommen!
Wenn ihr die Möglichkeit habt, würde ich mich riesig über das Teilen freuen, damit mehr Menschen auf ME/CFS und die dringend notwendige Forschung aufmerksam werden.
betterplace.org/de/fundraisi…
Paul Kaschner retweeted
Thrilled to announce $5B in funding for @IsomorphicLabs to accelerate our mission to Solve All Disease. We continue to make rapid progress reimagining drug design from first principles with AI to radically cut the time & cost to find new medicines. Gradually, then all at once...
Heute werde ich 23 und verbringe meinen Geburtstag wegen ME/CFS im Bett.
Viele von euch hier wissen leider selbst nur zu gut, was diese Erkrankung bedeutet.
Deshalb habe ich eine Spendenaktion für die Mecfs Research Foundation gestartet.
Und bereits jetzt sind über 1000€ innerhalb von einem Tag zusammengekommen!
Wenn ihr die Möglichkeit habt, würde ich mich riesig über das Teilen freuen, damit mehr Menschen auf ME/CFS und die dringend notwendige Forschung aufmerksam werden.
betterplace.org/de/fundraisi…
Paul Kaschner retweeted
🚨 Rumors I’ve been hearing. First OpenAI, then Anthropic and Google.
OPENAI
OpenAI is in a “safety lockdown” on its frontier work, tightening safety measures and containment protocols to avoid another Hugging Face incident. Training continues under much stricter controls.
Most gains come from video games and extreme sandboxes where models must complete clear, difficult long-term goals. These force planning, experimentation, recovery from failure, and strategy discovery, driving intelligence and creativity.
Public research roles match this: Synthetic RL uses self-play and simulators; agent teams focus on sustained execution and multi-model coordination; the Recursive Self-Improvement team builds systems that generate hypotheses, run experiments, and automate research.
The internal model is still training but nearing completion. The jump is enormous. narrow superintelligence in specific areas and overall AGI-level capability. Its ability to independently handle difficult work sets it apart.
ANTHROPIC
Model 5 is training and is expected to solve extremely hard problems in mathematics, biology, and other sciences. Anthropic is pushing models toward original research contributions, complex experiments, and sustained reasoning.
Fable 5.5 is Model 4. Fable 6 is Model 5; Opus 6 will be based on Model 5. Internal capabilities will feed the next generation of both families.
Training environments are central. The Universes team creates realistic situations with ambiguity, interruptions, and delayed consequences. The model must track the objective, recognize failing approaches, and find alternatives.
Public roles align: scaling RL across model size, compute, and task duration, plus environments for chip design, verification, and complex visual work. Multi-agent research explores how larger model teams can handle major codebases and difficult mathematics.
Model 5 advances Claude toward owning entire research and engineering workflows. Its role in developing the next model grows more significant.
Anthropic expects RSI March 2027.
GOOGLE
Google is pushing Gemini 5 for a major leap in reasoning, multimodal understanding, and independent long-horizon work. Potentially surpassing all labs by end of year. Carbon topping Artificial Analysis signaled the speed of their frontier progress.
World models are a key advantage. Genie 4 generates interactive environments that evolve with agent actions, enabling vast practice in planning, testing approaches, and learning from outcomes.
Public roles highlight persistent context and personalization: agents that retain useful information, adapt to users, and execute complex workflows without losing track.
Gemini 5 delivers stronger intelligence across text, images, audio, and video. Combined with Google’s infrastructure and simulation work, the scale is enormous.
All three labs are also advancing robotics and world models. intelligence that understands physical environments, predicts action consequences, and operates through real machines.
Meanwhile, their models increasingly contribute to creating their successors. Each improvement accelerates the research behind the next.
The next three months are going to be insane.
Could CAR-T cell therapy become a treatment option for ME/CFS in the future?
I found the video where Prof. Carmen Scheibenbogen discusses this promising research approach.
youtube.com/live/qDTpMhuUKgI…
1
2
249
Paul Kaschner retweeted
And just like that, we have our first clinical trial and pharma partnership in the works! We’re teaming up with Prolong to take the drug they developed for stroke and see if it can help our patients with ME/CFS by increasing oxygen delivery and blood flow. Thanks to the 200TB of data @chroniclebioAI has amassed, combined with our AI engine, we are identifying patients that show hypoxia so we can focus this trial on them.
This is really the promise of the company: understanding biology at such a deep level that we can match the exact right patients to the exact right drugs and massively increase the likelihood of success of trials.
Going from company creation to first trial in less than two years is pretty cool, but there’s still a lot of work to be done to prove we can improve patients’ lives. One step at a time.
Today we are announcing our first clinical trial. We're joining forces with Prolong to evaluate a potential treatment for post exertional malaise (PEM) in #MECFS.
Sanguinate is an investigational intravenous therapy designed to deliver oxygen to deprived tissues, support blood flow, and reduce inflammation. The planned multi-site Phase 2a trial will evaluate the drug in at least 40 patients with first dosing expected mid-2027.
Most trials enroll patients based on symptoms alone, but recruitment for this study will leverage ChronicleBio's proprietary endotype intelligence platform to identify patients whose disease biology suggests they may be more likely to respond.
We are excited for our partnership with Prolong and grateful for the patient community that drives everything we do.
Read More: chroniclebio.com/news/chroni…
Paul Kaschner retweeted
DARIO AMODEI WARNS AI WILL "OVERWHELM OUR ABILITY TO ADAPT"
"I know that the labor market is adaptable, right?"
"80% of people used to do farming"
"somewhere between a year and 5 years, it will overwhelm our ability to adapt"
adapting means getting accustomed to new rules, but in order to adapt, those rules first have to be defined. the issue with AI, and what makes it very different from other technological revolutions humanity has lived through, is the speed
i'm sure many of those reading this have been constantly adapting for the last two years, and yet the rules and reality just keep changing
at some point, it becomes exhausting, and you inevitably start questioning whether it's actually worth it. why spend time learning and adapting to something that might not even be relevant in 3-6 months?!
this is the challenge, and i guess this is exactly what Dario is talking about here
Was mir die letzten Tage auffällt: Die großen Pharma-CEOs sind alle begeistert von KI. Ob das dazu dient, den Börsenkurs zu verbessern? Oder wirklich, die eine Einschätzung bleibt abzuwarten. Aber auf alle Fälle schon mal sehr optimistisch für die Zukunft der Medizin.
Share Dario’s optimism. AI is changing the pace of medical progress, and in our industry time is life. Already seeing this across @Pfizer from discovery to manufacturing.
But as a scientist I also know how humbling biology can be. Identifying a promising molecule is just the beginning. The only way to know if a medicine is safe and effective is careful, rigorous testing in patients. That takes time, and needs experienced researchers and physicians in the loop.
The future I see is AI’s speed and scientific discipline working together. If we get that right, patients will benefit sooner than many expect.
1
64
Paul Kaschner retweeted
Share Dario’s optimism. AI is changing the pace of medical progress, and in our industry time is life. Already seeing this across @Pfizer from discovery to manufacturing.
But as a scientist I also know how humbling biology can be. Identifying a promising molecule is just the beginning. The only way to know if a medicine is safe and effective is careful, rigorous testing in patients. That takes time, and needs experienced researchers and physicians in the loop.
The future I see is AI’s speed and scientific discipline working together. If we get that right, patients will benefit sooner than many expect.
Paul Kaschner retweeted
Anyhow, in a little over a week we will see the release of The Last Human Game.
There will never be another AAA game at the scale of GTA 6 with no AI assistance. People will look back at an all-human game with a billion dollar budget and wonder how they did it, like the Pyramids
Die LIFT-Studie zu Pyridostigmin und Low-Dose Naltrexon bei ME/CFS und Long COVID wurde leider deutlich verlängert.
Primärer Studienabschluss ist jetzt September 2027. Eigentlich sollte dieser schon diesen September sein.
Damit verzögern sich mögliche Wirksamkeitsdaten um mindestens ein Jahr. Gründe sind bisher nicht bekannt.
2
73
Paul Kaschner retweeted
Unfortunately, it seems like the intelligence explosion will be happening in 2027 after all. Most of this post analyzes the new METR report (though other sources of evidence are also important). If they’re right, that implies an AI-driven acceleration that is 5 months ahead of the AI 2027 scenario. I am scared, sad, and angry. blog.aifutures.org/p/q35-tim…
Paul Kaschner retweeted
In one of my favorite experiments, I told Fable 5.2 to build an endless, hands-off journey through 3D worlds linked by portals.
25,000 agents created over 450 worlds inspired by books, art, places, all in code.
Play: portals-endless-worlds.netli…
Github: github.com/emollick/ai-porta…
Paul Kaschner retweeted
Southern Medical University researchers developed two nanobodies that block SARS-CoV-2 from fusing with human cells.
They neutralized live virus and reduced lung inflammation in mice by targeting a conserved region of spike.
Early but promising.
sciencedirect.com/science/ar…
Paul Kaschner retweeted
Since August, a team within METR has worked with internal Anthropic information to assess whether AI itself is accelerating AI development.
We are now publishing a report summarizing the results of this risk assessment.
Paul Kaschner retweeted
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-sp…
github.com/RikiyaT/lm_und…
huggingface.co/Qwen/Qwen3.5-0…
huggingface.co/RikiyaT/Qwen3.…
Paul Kaschner retweeted
8 of 12 people with autoimmune disease got a single infusion of CAR-T cells and went into lasting remission - years, no drugs.
With rituximab, the standard way of wiping out B cells, 0 of 15 did. A new preprint argues the gap comes down to one kind of cell - a B cell infected with EBV.🧵
Stanford researchers studied 27 autoimmune patients and found CD19 CAR T-cell therapy cleared EBV-infected, autoreactive B cells that resisted rituximab.
8 of 12 CAR T-cell patients achieved sustained drug-free remission.
biorxiv.org/content/10.64898…
Paul Kaschner retweeted
When you think about the Labs' weird approach to software (sudden changes, bad communication), it is worth remembering that a core belief is that these products are a temporary thing they have to do to raise money for a couple years until "true" AGI makes all products obsolete.
Die nächsten 2 Jahre werden wahrscheinlich die entscheidenden Jahre für die KI werden und die Arbeitswelt werden. Wenn man sich alleine den Fortschritt der KI seit Anfang dieses Jahres anschaut mag man sich nicht ausmalen was KI in 2 Jahren alles können wird.
ANTHROPIC: "we may imminently meet our Responsible Scaling Policy’s definition of 'dramatically accelerated progress attributable to AI.'"
"we do not anticipate being able to safely use Claude to fully autonomously design and build successive generations of frontier models"
107




















