DrDan
@drdanponders.bsky.social
PhD in AI (Edinburgh), ex-Google, ex-Alexa, serial failed entrepreneur, full time digital nomad for 5 years. Currently trying to on help researchers research faster to reverse aging.
Opinions are mine, but you're welcome to them
Opinions are mine, but you're welcome to them
Reposted by DrDan
As we can expect PubMed dysfunction more and more or to be swamped with fake journals I suggest you get acquainted to this nice alternative: europepmc.org
Europe PMCHome - Europe PMC
Europe PMC is an archive of life sciences journal literature.
europepmc.org
March 2, 2025 at 7:54 AM
As we can expect PubMed dysfunction more and more or to be swamped with fake journals I suggest you get acquainted to this nice alternative: europepmc.org
Reposted by DrDan
Can confirm. Former USAID here, and I’ve been approached twice in the last 30 days.
Your new allies: www.cnn.com/2025/02/28/p...
US intel shows Russia and China are attempting to recruit disgruntled federal employees, sources say | CNN Politics
Foreign adversaries including Russia and China have recently directed their intelligence services to ramp up recruiting of US federal employees working in national security, targeting those who have b...
www.cnn.com
March 1, 2025 at 4:52 PM
Can confirm. Former USAID here, and I’ve been approached twice in the last 30 days.
Reposted by DrDan
for all that is holy, what is this? how is it so good?
and who is sesame and why is their advanced voice mode 10 years ahead of OpenAI or anyone else?
www.sesame.com/research/cro...
and who is sesame and why is their advanced voice mode 10 years ahead of OpenAI or anyone else?
www.sesame.com/research/cro...
Crossing the uncanny valley of conversational voice
At Sesame, our goal is to achieve “voice presence”—the magical quality that makes spoken interactions feel real, understood, and valued.
www.sesame.com
March 1, 2025 at 8:48 PM
for all that is holy, what is this? how is it so good?
and who is sesame and why is their advanced voice mode 10 years ahead of OpenAI or anyone else?
www.sesame.com/research/cro...
and who is sesame and why is their advanced voice mode 10 years ahead of OpenAI or anyone else?
www.sesame.com/research/cro...
Reposted by DrDan
Silver lining:
Europe became a Union yesterday.
🇪🇺
Europe became a Union yesterday.
🇪🇺
March 1, 2025 at 10:04 AM
Silver lining:
Europe became a Union yesterday.
🇪🇺
Europe became a Union yesterday.
🇪🇺
Reposted by DrDan
18F staffers- who were summarily fired en masse overnight- issue statement expressing shock, bewilderment:
"18F was doing exactly the type of work that DOGE claims to want – yet we were eliminated."
"18F was doing exactly the type of work that DOGE claims to want – yet we were eliminated."
We're not done yet | 18F
18f.org
March 1, 2025 at 10:27 PM
18F staffers- who were summarily fired en masse overnight- issue statement expressing shock, bewilderment:
"18F was doing exactly the type of work that DOGE claims to want – yet we were eliminated."
"18F was doing exactly the type of work that DOGE claims to want – yet we were eliminated."
Reposted by DrDan
I will say again: The destruction of the CFPB—an agency that has brought in $20 billion over a little more than a decade despite having an annual budget that's less than what DOD has spent so far today—is the clearest sign none of this has anything to do with saving taxpayers money.
"I am preparing this declaration to ensure that the Court is aware of the imminent risk that twelve years’ of critical CFPB records, which belong to the public, will be irretrievably lost and cause serious and sweeping damage unless the Court takes action."
CFPB's former chief technologist issues an urgent warning to the court that several years' worth of sensitive data is about to be destroyed, with damaging consequences
storage.courtlistener.com/recap/gov.us...
storage.courtlistener.com/recap/gov.us...
February 16, 2025 at 6:49 PM
I will say again: The destruction of the CFPB—an agency that has brought in $20 billion over a little more than a decade despite having an annual budget that's less than what DOD has spent so far today—is the clearest sign none of this has anything to do with saving taxpayers money.
Reposted by DrDan
Government does important work — but not if DOGE can prevent it paulkrugman.substack.com/p/our-govern...
Our Government Is Experiencing a Rapid Unscheduled Disassembly
Musk is moving fast and breaking important things
paulkrugman.substack.com
February 17, 2025 at 11:06 AM
Government does important work — but not if DOGE can prevent it paulkrugman.substack.com/p/our-govern...
Reposted by DrDan
This cartoon by @mluckovich.bsky.social is getting banned on Twitter. Because of course it is.
February 17, 2025 at 12:26 AM
This cartoon by @mluckovich.bsky.social is getting banned on Twitter. Because of course it is.
Reposted by DrDan
This is what oligarchy looks like.
February 16, 2025 at 9:57 PM
This is what oligarchy looks like.
Reposted by DrDan
Increasingly large food.
All made by me with veo 2, the serious point is to note how impressive the "physics" of these models have become (also the consistency, like the spilled burrito across two different shots), there are some issues, obviously, but big advances.
All made by me with veo 2, the serious point is to note how impressive the "physics" of these models have become (also the consistency, like the spilled burrito across two different shots), there are some issues, obviously, but big advances.
January 8, 2025 at 9:01 PM
Increasingly large food.
All made by me with veo 2, the serious point is to note how impressive the "physics" of these models have become (also the consistency, like the spilled burrito across two different shots), there are some issues, obviously, but big advances.
All made by me with veo 2, the serious point is to note how impressive the "physics" of these models have become (also the consistency, like the spilled burrito across two different shots), there are some issues, obviously, but big advances.
Reposted by DrDan
“The effect may be considerable in a few centuries” is a really remarkable sentence here…
January 1, 2025 at 5:32 PM
“The effect may be considerable in a few centuries” is a really remarkable sentence here…
Y'all been asking, so I made my first proper youtube video to explain why o3 is a huge deal.
Like, what IS it? What does it all MEAN? Is your job obsolete already?
Find my best effort answers here :)
youtu.be/HM1XZ_3Uho8
Like, what IS it? What does it all MEAN? Is your job obsolete already?
Find my best effort answers here :)
youtu.be/HM1XZ_3Uho8
OpenAI o3 - AGI or hype?
YouTube video by DrDan ponders -- Daniel Duma
youtu.be
December 22, 2024 at 11:56 PM
Y'all been asking, so I made my first proper youtube video to explain why o3 is a huge deal.
Like, what IS it? What does it all MEAN? Is your job obsolete already?
Find my best effort answers here :)
youtu.be/HM1XZ_3Uho8
Like, what IS it? What does it all MEAN? Is your job obsolete already?
Find my best effort answers here :)
youtu.be/HM1XZ_3Uho8
Reposted by DrDan
THREAD
This Is The Story Of The The Pernicious Rise of AI-Generated Papers and their Online Impact
An Incomplete History Told In The Voice of Documentarian Adam Curtis
1/31
This Is The Story Of The The Pernicious Rise of AI-Generated Papers and their Online Impact
An Incomplete History Told In The Voice of Documentarian Adam Curtis
1/31
December 19, 2024 at 9:10 PM
THREAD
This Is The Story Of The The Pernicious Rise of AI-Generated Papers and their Online Impact
An Incomplete History Told In The Voice of Documentarian Adam Curtis
1/31
This Is The Story Of The The Pernicious Rise of AI-Generated Papers and their Online Impact
An Incomplete History Told In The Voice of Documentarian Adam Curtis
1/31
Reposted by DrDan
Just 10 days after o1's public debut, we’re thrilled to unveil the open-source version of the technique behind its success: scaling test-time compute
By giving models more "time to think," Llama 1B outperforms Llama 8B in math—beating a model 8x its size. The full recipe is open-source!
By giving models more "time to think," Llama 1B outperforms Llama 8B in math—beating a model 8x its size. The full recipe is open-source!
December 16, 2024 at 9:42 PM
Just 10 days after o1's public debut, we’re thrilled to unveil the open-source version of the technique behind its success: scaling test-time compute
By giving models more "time to think," Llama 1B outperforms Llama 8B in math—beating a model 8x its size. The full recipe is open-source!
By giving models more "time to think," Llama 1B outperforms Llama 8B in math—beating a model 8x its size. The full recipe is open-source!
Reposted by DrDan
It’s interesting that the emojis show the earth from three angles 🌎 🌍 🌏 but not the fourth.
Though 🔵 is a fair approximation of what it would look like.
People underestimate the Pacific.
Though 🔵 is a fair approximation of what it would look like.
People underestimate the Pacific.
December 6, 2024 at 3:52 PM
It’s interesting that the emojis show the earth from three angles 🌎 🌍 🌏 but not the fourth.
Though 🔵 is a fair approximation of what it would look like.
People underestimate the Pacific.
Though 🔵 is a fair approximation of what it would look like.
People underestimate the Pacific.
Reposted by DrDan
Been working on making it possible to connect Claude Desktop to a Cloudflare Worker using the new MCP spec, and it works!
You write some functions in a Worker, document them with JSDoc, and then they show up in Claude for you to use. It's kinda magical.
github.com/geelen/worke...
You write some functions in a Worker, document them with JSDoc, and then they show up in Claude for you to use. It's kinda magical.
github.com/geelen/worke...
GitHub - geelen/workers-mcp-server: Talk to a Cloudflare Worker from Claude Desktop!
Talk to a Cloudflare Worker from Claude Desktop! Contribute to geelen/workers-mcp-server development by creating an account on GitHub.
github.com
December 6, 2024 at 11:45 AM
Been working on making it possible to connect Claude Desktop to a Cloudflare Worker using the new MCP spec, and it works!
You write some functions in a Worker, document them with JSDoc, and then they show up in Claude for you to use. It's kinda magical.
github.com/geelen/worke...
You write some functions in a Worker, document them with JSDoc, and then they show up in Claude for you to use. It's kinda magical.
github.com/geelen/worke...
Reposted by DrDan
🧪 Ready or not, more medical AI is coming.
KLAS reports nearly 20% of healthcare orgs use AI, with larger systems (42%) leading adoption. Popular uses: neurology, breast cancer screening, lung nodule detection 🩺🖥️ #MLSky
KLAS reports nearly 20% of healthcare orgs use AI, with larger systems (42%) leading adoption. Popular uses: neurology, breast cancer screening, lung nodule detection 🩺🖥️ #MLSky
Up to half of medical organizations either already using or preparing to implement AI
With the number of FDA approved AI-enabled applications nearing 1,000, it’s clear the healthcare industry is ready to embrace the evolving technology.
healthimaging.com
December 6, 2024 at 4:11 PM
🧪 Ready or not, more medical AI is coming.
KLAS reports nearly 20% of healthcare orgs use AI, with larger systems (42%) leading adoption. Popular uses: neurology, breast cancer screening, lung nodule detection 🩺🖥️ #MLSky
KLAS reports nearly 20% of healthcare orgs use AI, with larger systems (42%) leading adoption. Popular uses: neurology, breast cancer screening, lung nodule detection 🩺🖥️ #MLSky
researcher: "we told the AI to be evil"
AI: *is evil*
researcher:
AI: *is evil*
researcher:
December 6, 2024 at 12:13 PM
researcher: "we told the AI to be evil"
AI: *is evil*
researcher:
AI: *is evil*
researcher:
Reposted by DrDan
Exponentially growing number of open-source AI models over the course of the past 30 months – from a few thousands to over 1 million and more
Interactive data viz: huggingface.co/spaces/huggi...
Interactive data viz: huggingface.co/spaces/huggi...
December 6, 2024 at 8:14 AM
Exponentially growing number of open-source AI models over the course of the past 30 months – from a few thousands to over 1 million and more
Interactive data viz: huggingface.co/spaces/huggi...
Interactive data viz: huggingface.co/spaces/huggi...
Reposted by DrDan
OpenAI in 2024:
“No AI for weapons or military”
“Do use our AI to make weapons to hurt yourself or others”
“Military is fine, but no AI for weapons”
“Sure put it on battlefield drones”
www.technologyreview.com/2024/12/04/1...
“No AI for weapons or military”
“Do use our AI to make weapons to hurt yourself or others”
“Military is fine, but no AI for weapons”
“Sure put it on battlefield drones”
www.technologyreview.com/2024/12/04/1...
OpenAI’s new defense contract completes its military pivot
A new partnership with Anduril, announced today, will deploy AI on the battlefield. It represents an overhaul of the company’s position in just a year.
www.technologyreview.com
December 4, 2024 at 11:43 PM
OpenAI in 2024:
“No AI for weapons or military”
“Do use our AI to make weapons to hurt yourself or others”
“Military is fine, but no AI for weapons”
“Sure put it on battlefield drones”
www.technologyreview.com/2024/12/04/1...
“No AI for weapons or military”
“Do use our AI to make weapons to hurt yourself or others”
“Military is fine, but no AI for weapons”
“Sure put it on battlefield drones”
www.technologyreview.com/2024/12/04/1...
Reposted by DrDan
From this article about chatbots that are designed to mimic human interaction: “Few behaviors trigger the anthropomorphic reflex as powerfully as the use of language” www.theverge.com/c/24300623/a...
What do you love when you fall for AI?
Inside the surprisingly meaningful, unexpectedly heartbreaking, and deeply confusing reality of AI relationships.
www.theverge.com
December 5, 2024 at 7:55 AM
From this article about chatbots that are designed to mimic human interaction: “Few behaviors trigger the anthropomorphic reflex as powerfully as the use of language” www.theverge.com/c/24300623/a...
Reposted by DrDan
The Discover feed is pretty terrible, and I am not sure how to guide the algorithm to make it better.
I know I can use various lists, but I really want a feed that includes serendipity, trends, and a broader set of topics than the one that I follow but isn't full of noise. How do you get that?
I know I can use various lists, but I really want a feed that includes serendipity, trends, and a broader set of topics than the one that I follow but isn't full of noise. How do you get that?
December 5, 2024 at 5:15 AM
The Discover feed is pretty terrible, and I am not sure how to guide the algorithm to make it better.
I know I can use various lists, but I really want a feed that includes serendipity, trends, and a broader set of topics than the one that I follow but isn't full of noise. How do you get that?
I know I can use various lists, but I really want a feed that includes serendipity, trends, and a broader set of topics than the one that I follow but isn't full of noise. How do you get that?
Last Thursday I gave my first invited talk at a university in many years: Politehnica Bucuresti is the STEM factory of Romania -- many of the students end up at a FAAMG / Magnificent Seven corp.
I had a room full of talented and driven students.
I had a room full of talented and driven students.
December 4, 2024 at 7:36 PM
Last Thursday I gave my first invited talk at a university in many years: Politehnica Bucuresti is the STEM factory of Romania -- many of the students end up at a FAAMG / Magnificent Seven corp.
I had a room full of talented and driven students.
I had a room full of talented and driven students.
Reposted by DrDan
Introducing 🧞Genie 2 🧞 - our most capable large-scale foundation world model, which can generate a diverse array of consistent worlds, playable for up to a minute. We believe Genie 2 could unlock the next wave of capabilities for embodied agents 🧠.
December 4, 2024 at 4:01 PM
Introducing 🧞Genie 2 🧞 - our most capable large-scale foundation world model, which can generate a diverse array of consistent worlds, playable for up to a minute. We believe Genie 2 could unlock the next wave of capabilities for embodied agents 🧠.