Aiden Barin
@abarin.bsky.social
data science @ uc health data warehouse.
Reposted by Aiden Barin
New AI Grand Rounds!
Dr. Karandeep Singh explains how early experiments with code led him to biomedical informatics, why gaps between paper performance and clinical reality must be confronted, and how governance committees weigh ethics and safety. Listen to the full episode: nejm.ai/ep34
Dr. Karandeep Singh explains how early experiments with code led him to biomedical informatics, why gaps between paper performance and clinical reality must be confronted, and how governance committees weigh ethics and safety. Listen to the full episode: nejm.ai/ep34
September 17, 2025 at 1:33 PM
New AI Grand Rounds!
Dr. Karandeep Singh explains how early experiments with code led him to biomedical informatics, why gaps between paper performance and clinical reality must be confronted, and how governance committees weigh ethics and safety. Listen to the full episode: nejm.ai/ep34
Dr. Karandeep Singh explains how early experiments with code led him to biomedical informatics, why gaps between paper performance and clinical reality must be confronted, and how governance committees weigh ethics and safety. Listen to the full episode: nejm.ai/ep34
Reposted by Aiden Barin
August 9, 2025 at 5:02 AM
Reposted by Aiden Barin
"All the data in the world will not make a difference, unless there are people ready to use it."
— Atul Butte, MD, PHD, FACMI
2024 Morris F. Collen Award Winner
— Atul Butte, MD, PHD, FACMI
2024 Morris F. Collen Award Winner
November 16, 2024 at 1:58 AM
"All the data in the world will not make a difference, unless there are people ready to use it."
— Atul Butte, MD, PHD, FACMI
2024 Morris F. Collen Award Winner
— Atul Butte, MD, PHD, FACMI
2024 Morris F. Collen Award Winner
Reposted by Aiden Barin
Proud to be the first chief AI officer to have their job replaced by an AI.
May 29, 2025 at 3:22 AM
Proud to be the first chief AI officer to have their job replaced by an AI.
Reposted by Aiden Barin
The annual UC Health conference is a great place to connect with colleagues, both from Michigan and across the UC system.
Really enjoyed learning about:
- Opportunities/challenges with automation and AI
- UCSF’s locally trained GPT
- UCLA/Riverside/Davis efforts around health equity
Really enjoyed learning about:
- Opportunities/challenges with automation and AI
- UCSF’s locally trained GPT
- UCLA/Riverside/Davis efforts around health equity
April 25, 2025 at 2:23 PM
The annual UC Health conference is a great place to connect with colleagues, both from Michigan and across the UC system.
Really enjoyed learning about:
- Opportunities/challenges with automation and AI
- UCSF’s locally trained GPT
- UCLA/Riverside/Davis efforts around health equity
Really enjoyed learning about:
- Opportunities/challenges with automation and AI
- UCSF’s locally trained GPT
- UCLA/Riverside/Davis efforts around health equity
Reposted by Aiden Barin
Llama 4 is released - all versions are mixture of experts models, so the number of “active parameters” is much lower than the number of total parameters (similar to Mixtral).
RIP semver - we now have a Scout, Maverick, and Behemoth.
Biggest deal: 10M token context
ai.meta.com/blog/llama-4...
RIP semver - we now have a Scout, Maverick, and Behemoth.
Biggest deal: 10M token context
ai.meta.com/blog/llama-4...
The Llama 4 herd: The beginning of a new era of natively multimodal AI innovation
We’re introducing Llama 4 Scout and Llama 4 Maverick, the first open-weight natively multimodal models with unprecedented context support and our first built using a mixture-of-experts (MoE) architect...
ai.meta.com
April 6, 2025 at 12:22 AM
Llama 4 is released - all versions are mixture of experts models, so the number of “active parameters” is much lower than the number of total parameters (similar to Mixtral).
RIP semver - we now have a Scout, Maverick, and Behemoth.
Biggest deal: 10M token context
ai.meta.com/blog/llama-4...
RIP semver - we now have a Scout, Maverick, and Behemoth.
Biggest deal: 10M token context
ai.meta.com/blog/llama-4...
Reposted by Aiden Barin
Transparency is key to advancing #AI in #healthcare. In a letter to the White House’s AI Action Plan, @kdpsingh.bsky.social @chrislonghurst.bsky.social @atulbutte.bsky.social call for clearer guidelines to help health systems make informed decisions.
Read⤵️
healthinnovation.ucsd.edu/news/respons...
Read⤵️
healthinnovation.ucsd.edu/news/respons...
March 17, 2025 at 10:25 PM
Transparency is key to advancing #AI in #healthcare. In a letter to the White House’s AI Action Plan, @kdpsingh.bsky.social @chrislonghurst.bsky.social @atulbutte.bsky.social call for clearer guidelines to help health systems make informed decisions.
Read⤵️
healthinnovation.ucsd.edu/news/respons...
Read⤵️
healthinnovation.ucsd.edu/news/respons...
Reposted by Aiden Barin
We’re excited about @ucsdhealth.bsky.social launch of 2 HIPAA-compliant GenAI tools: MS Copilot Chat and Doximity GPT! These tools will help all team members streamline their work and focus more on #patientcare.
JCHI is here to support innovative applications!
@microsoft.com @doximity.bsky.social
JCHI is here to support innovative applications!
@microsoft.com @doximity.bsky.social
February 12, 2025 at 6:22 PM
We’re excited about @ucsdhealth.bsky.social launch of 2 HIPAA-compliant GenAI tools: MS Copilot Chat and Doximity GPT! These tools will help all team members streamline their work and focus more on #patientcare.
JCHI is here to support innovative applications!
@microsoft.com @doximity.bsky.social
JCHI is here to support innovative applications!
@microsoft.com @doximity.bsky.social
Reposted by Aiden Barin
Meta’s Llama 3.3 is multilingual. It technically supports English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai, but it was trained on many others, including Punjabi!
Me: What’s your name?
Llama 3.3: My name is “Language Model,” often called Llama.
huggingface.co/meta-llama/L...
Me: What’s your name?
Llama 3.3: My name is “Language Model,” often called Llama.
huggingface.co/meta-llama/L...
December 8, 2024 at 5:41 AM
Meta’s Llama 3.3 is multilingual. It technically supports English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai, but it was trained on many others, including Punjabi!
Me: What’s your name?
Llama 3.3: My name is “Language Model,” often called Llama.
huggingface.co/meta-llama/L...
Me: What’s your name?
Llama 3.3: My name is “Language Model,” often called Llama.
huggingface.co/meta-llama/L...
Reposted by Aiden Barin
Whoa! @Amazon to invest an additional $4 billion in AI startup @Anthropic!
Amazon to invest an additional $4 billion in AI startup Anthropic
Amazon is investing an additional $4 billion in the artificial intelligence startup Anthropic as major technology companies rush to fund generative AI.
buff.ly
December 5, 2024 at 8:47 AM
Whoa! @Amazon to invest an additional $4 billion in AI startup @Anthropic!