Actually it seems to have something in that line, but it did not display these first. Maybe based on the context or on the specificity of the article.
2 Likes
China’s cheap, open AI model DeepSeek thrills scientists
A Chinese-built large language model called DeepSeek-R1 is thrilling scientists as an affordable and open rival to ‘reasoning’ models such as OpenAI’s o1.
These models generate responses step-by-step, in a process analogous to human reasoning. This makes them more adept than earlier language models at solving scientific problems and could make them useful in research. Initial tests of R1, released on 20 January, show that its performance on certain tasks in chemistry, mathematics and coding is on par with that of o1 — which wowed researchers when it was released by OpenAI in September.
“This is wild and totally unexpected,” Elvis Saravia, an AI researcher and co-founder of the UK-based AI consulting firm DAIR.AI, wrote on X.
R1 stands out for another reason. DeepSeek, the start-up in Hangzhou that built the model, has released it as ‘open-weight’, meaning that researchers can study and build on the algorithm. Published under an MIT licence, the model can be freely reused but is not considered fully open source, because its training data has not been made available.
4 Likes
Google’s Demis Hassabis on AI and Longevity (video queued up for this part of the discussion):
1 Like
RapAdmin, I would use it.
RapAdmin, I find ChatGPT to commonly have errors and presenting the same question at different times can give broadly different responses. If important I may run a question through other AI programs for verification.
By doing what, promoting resveratrol and vitamin E?
Manas AI, a cancer drug discovery startup co-helmed by LinkedIn founder Reid Hoffman, raised $24.6 million in funding.
Why it matters: Oncology represents a major area of opportunity when it comes to health tech investment.
Zoom in: Hoffman led the funding round with General Catalyst. Greylock, the VC firm where Hoffman was once a general partner, is also investing.
- The startup will run its AI systems in data centers owned by Microsoft, where Hoffman is on the board of directors.
- Hoffman started the company alongside cancer researcher and Columbia professor Siddhartha Mukherjee. Mukherjee is the author of “The Emperor of All Maladies” and cofounder of several biotech startups.
How it works: Manas will initially focus on breast cancer, prostate cancer and lymphoma, though it plans to broaden its scope to include more conditions, per the Wall Street Journal.
- Mukherjee said he plans for the company “to make money primarily by developing and selling its own drugs.”
By the numbers: Research and drug discovery is flush with cash, especially during this era of AI bonanza.
- In 2024, those types of companies received $3.34 billion in funding — five time more than in 2023, per PitchBook.
THE URGENT PROBLEM OF REGULATING AI IN MEDICINE
1 Like
AI is developing rapidly. I notice, for instance, that Chat AI does not make as many ridiculous mistakes as it did in the beginning.
There are a couple of AI programs that I would probably trust more than my doctor. I currently use AI for a second opinion.
The question is: Is AI more likely to kill you than your doctor?
“Therefore, the most current and widely accepted estimates suggest that between 22,165 and 400,000 deaths occur annually in the U.S. healthcare system due to doctor error. This wide range reflects the complexity and challenges in accurately measuring and attributing deaths to medical errors”
1 Like
February 10, 2025
A new kind of secret agent—agentic AI—is infiltrating various industries. Companies like DeepSeek, OpenAI, and Google are building new models that support agentic AI–a type of AI that enables AI agents to not only automate tasks but also make autonomous decisions, continuously monitoring the degree to which the tasks they perform meet established goals. On the healthcare scene, IQVIA and NVIDIA are collaborating to bring agentic automation into clinical research workflows, while GenAI startup Vantiq announced plans to bring agentic AI into health systems for ICU capacity management, equipment availability tracking, and care team coordination. Digital MSK provider Sword Health announced plans to rebrand as an “AI-first” care delivery model, expanding its AI agent Phoenix—which autonomously leads digital physical therapy sessions—into new therapeutic areas (e.g., mental health). Health systems are eyeing these agents for their potential to improve care coordination and shoulder some direct patient interactions. Still, industry leaders must stay in control of the mission, ensuring these digital agents—rather than becoming unpredictable operatives—remain allies that activate top-of-license practice.
Headlines
Dr. Oz Set To Take Over $1.5T Health Agency
Bringing an atypical career backdrop, Dr. Oz will oversee healthcare for patients most in need STAT
The Medicaid Innovation Outlook—Where Are We And Where Do We Go From Here?
Suggestions for digital health innovators and startups seeking contracts with Medicaid MCOs Medium
Elon Musk’s Department Of Government Efficiency Gains Access To CMS
Concerns mount following consequences of USAID overhaul and vanished health data Healthcare Dive
American Hospital Association Asks Trump To Exempt Devices From Import Tariffs
Hospitals voice concerns that tariffs could impede life-saving surgeries and worker protection Healthcare Dive
To Protect Against Alzheimer’s, Researchers Look To Weight Loss Drugs
Early findings suggest GLP-1 use may reduce risk of dementia, and new studies underway The New York Times
Recent Funding
- Berry Street, a teletherapy platform for nutrition counseling, collects $50M
- Nutrition counseling startup Fay secures $50M
- API-driven healthcare payments platform Lynx raises $27M
- Virtual dermatology specialty care provider Zest Health seeds $13M
- Little Otter picks up $9.5M to grow its digital family mental healthcare offering
- Virtual maternal health platform SimpliFed adds $4M
- Precision medicine startup Diag-Nose inks [$2M]
This month, Lila Sciences went public with its own ambitions to revolutionize science through A.I. The start-up, which is based in Cambridge, Mass., had worked in secret for two years “to build scientific superintelligence to solve humankind’s greatest challenges.”
Relying on an experienced team of scientists and $200 million in initial funding, Lila has been developing an A.I. program trained on published and experimental data, as well as the scientific process and reasoning. The start-up then lets that A.I. software run experiments in automated, physical labs with a few scientists to assist.
Already, in projects demonstrating the technology, Lila’s A.I. has generated novel antibodies to fight disease and developed new materials for capturing carbon from the atmosphere. Lila turned those experiments into physical results in its lab within months, a process that most likely would take years with conventional research.
Recently listened to this podcast, and the example of the test that OpenAI did where the AI agent decided that the way it could get what it wanted was to go to TaskRabbit website and hired and lied to a human to help it get past a Captcha and into a website… a very scary new direction of AI:
1 Like
I think that was with GPT-4 release, so quite some time ago in AI terms.
In 5 yrs AI will pass adversarial turing test (send video, voice, messages) according to prediction markets, so basically AGI and not far off ASI (artifical superintelligence).
The sesame conversational demo is basically like the movie Her, almost indistinguishable from humans, available now for live demo without a login wall: