Dag Hanoa
banner
daghanoa.bsky.social
Dag Hanoa
@daghanoa.bsky.social
Generative AI and deep learning in healthcare main interest. MD psychiatrist. Winter enthusiast.
The Human Therapist isn’t dead yet: New Research: Human Therapists Surpass ChatGPT in Delivering Cognitive Behavioral Therapy APA 2025

www.psychiatry.org/news-room/ne...
New Research: Human Therapists Surpass ChatGPT in Delivering Cognitive Behavioral Therapy
New research presented today at the American Psychiatric Association’s Annual Meeting compared an AI therapist and a human therapist based on their delivery of text-based cognitive behavioral therapy ...
www.psychiatry.org
May 18, 2025 at 7:17 PM
AGI Pioneer Peter Voss has just dropped one of the most profound pieces on the future of AI. aigo.ai/agi-from-fir...
AGI, from First Principles Aigo.ai - The Company with a Direct Path to AGI
What is AGI? In short, Artificial General Intelligence is the name given to computer systems that replicate the overall cognitive abilities of a smart human. In particular, to capture the unique menta...
aigo.ai
May 7, 2025 at 3:18 PM
Reposted by Dag Hanoa
"Scientists discover ancient brain-cleaning system that may break down in Alzheimer’s disease"

an interesting read!

I don't know enough sciency stuff to fully understand but it does make me wonder how this research might also relate to #ME/CFS and even #MS

#MedSky

www.psypost.org/scientist-di...
May 4, 2025 at 12:05 AM
Reposted by Dag Hanoa
Still Confused About the Types of Language Models?

From GPT-style models to advanced multilingual systems, Large Language Models (LLMs) come in various architectures, each suited for specific tasks. Here’s a breakdown of the 6 main types: 1. Autoregressive Models Generate text sequentially…
Still Confused About the Types of Language Models?
From GPT-style models to advanced multilingual systems, Large Language Models (LLMs) come in various architectures, each suited for specific tasks. Here’s a breakdown of the 6 main types: 1. Autoregressive Models Generate text sequentially (word-by-word) based on prior context. Example: GPT-3. - Use Case: Text generation, chatbots. - Limitation: Can produce repetitive outputs. 2. Transformer-Based Models Leverage self-attention mechanisms to process long-range dependencies.
undercodetesting.com
May 1, 2025 at 6:50 PM
Reposted by Dag Hanoa
Large language models are less effective at clinical prediction tasks than locally trained machine learning models academic.oup.com/jamia/advanc... #llms #machinelearning #MLSky (not surprising but important to document)
Large language models are less effective at clinical prediction tasks than locally trained machine learning models
AbstractObjectives. To determine the extent to which current large language models (LLMs) can serve as substitutes for traditional machine learning (ML) as
academic.oup.com
April 8, 2025 at 6:17 PM
Reposted by Dag Hanoa
We’ve been blessed with another Andrej Karpathy video 🥹 youtu.be/7xTGNNLPyMI
Deep Dive into LLMs like ChatGPT
YouTube video by Andrej Karpathy
youtu.be
February 5, 2025 at 6:35 PM
Reposted by Dag Hanoa
Happy Sunday blue sky! 3 GitHub Repos to learn more about #AI and #ML

1. Neural Networks: Zero to Hero from Andrej Karpathy.
lnkd.in/erA-Gerp

2. Machine Learning Andrew Ng’s lectures
lnkd.in/ebtWemWD

3. Deep Learning Specialization - Andrew Ng’s lectures & practice problems
lnkd.in/e2RECCSQ
January 19, 2025 at 9:06 PM
Reposted by Dag Hanoa
The human side of AI seems to be our biggest bottleneck: understanding the potential, drawbacks, threats, ethical implications, and limitations of models as they continue to advance.

Ethan Mollick has a great perspective on this today:
Prophecies of the Flood
What to make of the statements of the AI labs?
www.oneusefulthing.org
January 10, 2025 at 6:28 PM
Reposted by Dag Hanoa
If you haven’t found your people on #bluesky, here’s a readymade community with 4 starter packs of psychiatrists, psychologists, mental health workers, writers & thinkers, people with lived experience, & policy folks

go.bsky.app/C36fgiD

go.bsky.app/3UH1pKc

go.bsky.app/CVJ1byu

go.bsky.app/KYFiVUa
December 31, 2024 at 9:03 AM
The use of collective intelligence methods to synthesize differential diagnoses, combining the responses of different LLMs, achieves three of the necessary steps toward advancing LLMs as a diagnostic support tool: demonstrating sufficiently high diagnostic accuracy”

ai.nejm.org/doi/full/10....
Combining Multiple Large Language Models Improves Diagnostic Accuracy
Although large language models (LLMs), such as OpenAI GPT-4 or Google PaLM 2, are proposed as viable diagnostic support tools or even spoken of as replacements for “curbside consults,” past studies...
ai.nejm.org
December 16, 2024 at 8:08 PM