Michael
@michaelbujard.bsky.social
science | engineering | technology | mathematics | analytics | data | software | other | don't remember | random
Pinned
Michael
@michaelbujard.bsky.social
· Nov 8
Pope Leo XIV on X: "The world needs honest and courageous entrepreneurs and communicators who care for the common good. We sometimes hear the saying: “Business is business!” In reality, it is not so. No one is absorbed by an organization to the point of becoming a mere cog or a simple function. Nor" / X
The world needs honest and courageous entrepreneurs and communicators who care for the common good. We sometimes hear the saying: “Business is business!” In reality, it is not so. No one is absorbed by an organization to the point of becoming a mere cog or a simple function. Nor
x.com
I want to pin this to my wall to read every day
I hope Pope Leo gets a Bluesky account
It's a pain to switch between platforms because people are too divided to use just one.
#Family
#Work
#Motivation
x.com/Pontifex/sta...
x.com/Pontifex/sta...
I hope Pope Leo gets a Bluesky account
It's a pain to switch between platforms because people are too divided to use just one.
#Family
#Work
#Motivation
x.com/Pontifex/sta...
x.com/Pontifex/sta...
So funny but a critical issue in some contexts, since we believe it's the ordinary doorway to eternal life! Maybe the equivalent today:
Someone at an audience with the Pope asks: coffee/energy drinks are so abundant that, if someone in an emergency wanted to receive the Sacrament before death,
1/N
Someone at an audience with the Pope asks: coffee/energy drinks are so abundant that, if someone in an emergency wanted to receive the Sacrament before death,
1/N
Another beer post for the evening: In 1241, Pope Gregory IX informed Archbishop Sigurd Eindridesson (r.1231-52) that children who are baptised in beer can not be considered to have been properly baptised.
December 20, 2025 at 3:00 PM
So funny but a critical issue in some contexts, since we believe it's the ordinary doorway to eternal life! Maybe the equivalent today:
Someone at an audience with the Pope asks: coffee/energy drinks are so abundant that, if someone in an emergency wanted to receive the Sacrament before death,
1/N
Someone at an audience with the Pope asks: coffee/energy drinks are so abundant that, if someone in an emergency wanted to receive the Sacrament before death,
1/N
Reposted by Michael
Built a 2.5MB image classifier that runs in the browser in an evening with Claude Code.
I used a dataset I labelled in 2022 and left on @hf.co for 3 years 😬.
It finds illustrated pages in historical books. No server. No GPU.
I used a dataset I labelled in 2022 and left on @hf.co for 3 years 😬.
It finds illustrated pages in historical books. No server. No GPU.
December 19, 2025 at 12:08 PM
Built a 2.5MB image classifier that runs in the browser in an evening with Claude Code.
I used a dataset I labelled in 2022 and left on @hf.co for 3 years 😬.
It finds illustrated pages in historical books. No server. No GPU.
I used a dataset I labelled in 2022 and left on @hf.co for 3 years 😬.
It finds illustrated pages in historical books. No server. No GPU.
Reposted by Michael
Connect your agents directly to Chrome DevTools. 🛠️
The new Chrome DevTools MCP server allows agents to inspect the DOM and debug console errors in real-time using Gemini CLI or any compatible agent.
Check out the documentation on GitHub → goo.gle/CDT-MCP
The new Chrome DevTools MCP server allows agents to inspect the DOM and debug console errors in real-time using Gemini CLI or any compatible agent.
Check out the documentation on GitHub → goo.gle/CDT-MCP
December 20, 2025 at 4:00 AM
Connect your agents directly to Chrome DevTools. 🛠️
The new Chrome DevTools MCP server allows agents to inspect the DOM and debug console errors in real-time using Gemini CLI or any compatible agent.
Check out the documentation on GitHub → goo.gle/CDT-MCP
The new Chrome DevTools MCP server allows agents to inspect the DOM and debug console errors in real-time using Gemini CLI or any compatible agent.
Check out the documentation on GitHub → goo.gle/CDT-MCP
Hey, is this just a small thing, or is this important?
Cognitive Kernel Networks whitepaper, John P. Alioto.
Probably not on Bluesky. If he is, I'm slightly disappointed.
github.com/jpalioto/ckn...
Cognitive Kernel Networks whitepaper, John P. Alioto.
Probably not on Bluesky. If he is, I'm slightly disappointed.
github.com/jpalioto/ckn...
github.com
December 13, 2025 at 12:30 AM
Hey, is this just a small thing, or is this important?
Cognitive Kernel Networks whitepaper, John P. Alioto.
Probably not on Bluesky. If he is, I'm slightly disappointed.
github.com/jpalioto/ckn...
Cognitive Kernel Networks whitepaper, John P. Alioto.
Probably not on Bluesky. If he is, I'm slightly disappointed.
github.com/jpalioto/ckn...
I am preparing to take a statistics exam.
Can you please recommend a textbook?
A no-nonsense, dry, basic Stats 101 textbook that I can find for pennies on the dollar, if possible?
I haven't had one up to this point. I can't tell you why.
Can you please recommend a textbook?
A no-nonsense, dry, basic Stats 101 textbook that I can find for pennies on the dollar, if possible?
I haven't had one up to this point. I can't tell you why.
December 9, 2025 at 2:01 AM
I am preparing to take a statistics exam.
Can you please recommend a textbook?
A no-nonsense, dry, basic Stats 101 textbook that I can find for pennies on the dollar, if possible?
I haven't had one up to this point. I can't tell you why.
Can you please recommend a textbook?
A no-nonsense, dry, basic Stats 101 textbook that I can find for pennies on the dollar, if possible?
I haven't had one up to this point. I can't tell you why.
Reposted by Michael
A modest proposal that no one should be allowed to use “positivism” — and certainly not equate it with belief in “objectivity” — without adding a footnote where they explain whether Gelman & Hennig “Beyond subjective and objective in statistics” is also positivist.
![Beyond subjective and objective in statistics
Andrew Gelman
Columbia University, New York, USA
and Christian Hennig
University College London, UK
[Read before The Royal Statistical Society on Wednesday, April 12th, 2017, Professor P. J.
Diggle in the Chair]
Summary. Decisions in statistical data analysis are often justified, criticized or avoided by using concepts of objectivity and subjectivity. We argue that the words 'objective' and 'sub-jective' in statistics discourse are used in a mostly unhelpful way, and we propose to replace each of them with broader collections of attributes, with objectivity replaced by transparency, consensus, impartiality and correspondence to observable reality, and subjectivity replaced by awareness of multiple perspectives and context dependence. Together with stability, these make up a collection of virtues that we think is helpful in discussions of statistical foundations and practice. The advantage of these reformulations is that the replacement terms do not oppose each other and that they give more specific guidance about what statistical science strives to achieve. Instead of debating over whether a given statistical method is subjective or objective (or normatively debating the relative merits of subjectivity and objectivity in statistical practice), we can recognize desirable attributes such as transparency and acknowledgement of multiple perspectives as complementary goals. We demonstrate the implications of our proposal with recent applied examples from pharmacology, election polling and socio-economic stratification.
The aim of the paper is to push users and developers of statistical methods towards more effective use of diverse sources of information and more open acknowledgement of assumptions and goals.
Keywords: Bayesian; Frequentist; Good practice; Philosophy of statistics; Virtues](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:565ebob5f6hw33hjdkxty6qj/bafkreihjbl5gd2xodc6xuxmqdmw4ph3q5s6uaitbx7i6dxa4gmrfzhnyle@jpeg)
December 6, 2025 at 1:03 PM
A modest proposal that no one should be allowed to use “positivism” — and certainly not equate it with belief in “objectivity” — without adding a footnote where they explain whether Gelman & Hennig “Beyond subjective and objective in statistics” is also positivist.
Jon Krohn is one of the most generous techies out there, in my opinion.
He has a good podcast.
If you like podcasts, and if you are a data professional
Or someone who likes podcasts
Or just want to learn how to do 1:1 video sessions right
Then you might check out:
www.youtube.com/@SuperDataSc...
He has a good podcast.
If you like podcasts, and if you are a data professional
Or someone who likes podcasts
Or just want to learn how to do 1:1 video sessions right
Then you might check out:
www.youtube.com/@SuperDataSc...
Super Data Science: ML & AI Podcast with Jon Krohn
The latest machine learning, A.I., and data career topics from across both academia and industry are brought to you by host Dr. Jon Krohn on Super Data Science, the most listened-to podcast in the ind...
www.youtube.com
December 4, 2025 at 12:15 AM
Jon Krohn is one of the most generous techies out there, in my opinion.
He has a good podcast.
If you like podcasts, and if you are a data professional
Or someone who likes podcasts
Or just want to learn how to do 1:1 video sessions right
Then you might check out:
www.youtube.com/@SuperDataSc...
He has a good podcast.
If you like podcasts, and if you are a data professional
Or someone who likes podcasts
Or just want to learn how to do 1:1 video sessions right
Then you might check out:
www.youtube.com/@SuperDataSc...
every time i open bluesky the wish to read 200X faster becomes stronger
November 30, 2025 at 3:52 AM
every time i open bluesky the wish to read 200X faster becomes stronger
Reposted by Michael
Phil Foden's opener against Leeds was the first Premier League goal this season to be scored in the opening minute of a match.
November 29, 2025 at 3:14 PM
Phil Foden's opener against Leeds was the first Premier League goal this season to be scored in the opening minute of a match.
Reposted by Michael
According to the authors, ROOT outperforms Muon and Adam variants with faster convergence, higher final performance, and greater stability.
Paper: arxiv.org/abs/2511.20626
Repo: github.com/huawei-noah/...
Paper: arxiv.org/abs/2511.20626
Repo: github.com/huawei-noah/...
ROOT: Robust Orthogonalized Optimizer for Neural Network Training
The optimization of large language models (LLMs) remains a critical challenge, particularly as model scaling exacerbates sensitivity to algorithmic imprecision and training instability. Recent advance...
arxiv.org
November 29, 2025 at 3:22 PM
According to the authors, ROOT outperforms Muon and Adam variants with faster convergence, higher final performance, and greater stability.
Paper: arxiv.org/abs/2511.20626
Repo: github.com/huawei-noah/...
Paper: arxiv.org/abs/2511.20626
Repo: github.com/huawei-noah/...
Reposted by Michael
Huawei's Robust Orthogonalized Optimizer (ROOT), which brings two layers of robustness:
- Dimension-robust orthogonalization via adaptive Newton iterations with size-aware coefficients
- Optimization-robust updates using proximal methods that dampen harmful outliers while preserving useful gradient
- Dimension-robust orthogonalization via adaptive Newton iterations with size-aware coefficients
- Optimization-robust updates using proximal methods that dampen harmful outliers while preserving useful gradient
November 29, 2025 at 3:22 PM
Huawei's Robust Orthogonalized Optimizer (ROOT), which brings two layers of robustness:
- Dimension-robust orthogonalization via adaptive Newton iterations with size-aware coefficients
- Optimization-robust updates using proximal methods that dampen harmful outliers while preserving useful gradient
- Dimension-robust orthogonalization via adaptive Newton iterations with size-aware coefficients
- Optimization-robust updates using proximal methods that dampen harmful outliers while preserving useful gradient
Reposted by Michael
As someone who likes change — when it's slow enough to understand — I've been cheered up by Ilya's affirmation that scaling won't solve everything, and by recent signs that business adoption is slowing (Economist article below).
November 29, 2025 at 4:07 PM
As someone who likes change — when it's slow enough to understand — I've been cheered up by Ilya's affirmation that scaling won't solve everything, and by recent signs that business adoption is slowing (Economist article below).
Reposted by Michael
#AcademicSky I gave up with my quest to follow all Bioinformaticians/ Bioinformatics Scientists on #Bluesky as moderation doesn't like people who follow too many accounts compared to followers. I had to cull mine to 43k
We do understand that you want to follow—and be followed by—all the bioinformaticians on Bluesky, but we’re afraid your post might drown in a… blue sea! 😉 The hashtag AcademicSky will make your post appear in the feeds Academia and AcademicSky at the same time. Good luck in your quest!
November 29, 2025 at 6:02 PM
#AcademicSky I gave up with my quest to follow all Bioinformaticians/ Bioinformatics Scientists on #Bluesky as moderation doesn't like people who follow too many accounts compared to followers. I had to cull mine to 43k
Reposted by Michael
Headline: "AI can replace 11.7% of workforce"
Actual study: Anthropic paid MIT to use a "labor simulation tool" that said 11.7% of TASKS could be done by AI
Actual study: Anthropic paid MIT to use a "labor simulation tool" that said 11.7% of TASKS could be done by AI
MIT study finds AI can already replace 11.7% of U.S. workforce
Artificial intelligence can already replace 11.7% of the U.S. labor market, across finance, health care and professional services, according to MIT's study.
www.cnbc.com
November 29, 2025 at 9:57 PM
Headline: "AI can replace 11.7% of workforce"
Actual study: Anthropic paid MIT to use a "labor simulation tool" that said 11.7% of TASKS could be done by AI
Actual study: Anthropic paid MIT to use a "labor simulation tool" that said 11.7% of TASKS could be done by AI
Reposted by Michael
“.. Three years into the generative-ai wave, demand for the technology looks surprisingly flimsy.
“.. the employment-weighted share of Americans using ai at work has fallen by a percentage point, and now sits at 11% ..”
@economist.com
www.economist.com/finance-and-...
“.. the employment-weighted share of Americans using ai at work has fallen by a percentage point, and now sits at 11% ..”
@economist.com
www.economist.com/finance-and-...
November 29, 2025 at 9:35 PM
“.. Three years into the generative-ai wave, demand for the technology looks surprisingly flimsy.
“.. the employment-weighted share of Americans using ai at work has fallen by a percentage point, and now sits at 11% ..”
@economist.com
www.economist.com/finance-and-...
“.. the employment-weighted share of Americans using ai at work has fallen by a percentage point, and now sits at 11% ..”
@economist.com
www.economist.com/finance-and-...
Reposted by Michael
Three years into the generative-AI wave, demand for the technology seems surprisingly flimsy
Investors expect AI use to soar. That’s not happening
Recent surveys point to flatlining business adoption
econ.st
November 29, 2025 at 9:20 PM
Three years into the generative-AI wave, demand for the technology seems surprisingly flimsy
Reposted by Michael
The original.
The old sailing ship, hero of the Napoleonic wars, glowing in the sunset, is towed to the breakers by a coal powered tug belching black smoke. So there is a contrast between old and modern ships, with the idea that the modern successor is in some way fallen from the original in form.
The old sailing ship, hero of the Napoleonic wars, glowing in the sunset, is towed to the breakers by a coal powered tug belching black smoke. So there is a contrast between old and modern ships, with the idea that the modern successor is in some way fallen from the original in form.
November 30, 2025 at 2:40 AM
The original.
The old sailing ship, hero of the Napoleonic wars, glowing in the sunset, is towed to the breakers by a coal powered tug belching black smoke. So there is a contrast between old and modern ships, with the idea that the modern successor is in some way fallen from the original in form.
The old sailing ship, hero of the Napoleonic wars, glowing in the sunset, is towed to the breakers by a coal powered tug belching black smoke. So there is a contrast between old and modern ships, with the idea that the modern successor is in some way fallen from the original in form.
There is something otherworldly and apocalyptic about the optics of this historic meeting between Leo XIV and Ecumenical Patriarch Bartholomew. The image itself is like footage from another planet, and the accompanying music is absolutely alien, but it's from this world...
🌏
youtu.be/JT0jPea8XoQ?...
🌏
youtu.be/JT0jPea8XoQ?...
November 29, 2025 at 1:19 PM
There is something otherworldly and apocalyptic about the optics of this historic meeting between Leo XIV and Ecumenical Patriarch Bartholomew. The image itself is like footage from another planet, and the accompanying music is absolutely alien, but it's from this world...
🌏
youtu.be/JT0jPea8XoQ?...
🌏
youtu.be/JT0jPea8XoQ?...
Reposted by Michael
Straight from the horse's mouth:
“These models somehow just generalize dramatically worse than people. It's a very fundamental thing.”
“These models somehow just generalize dramatically worse than people. It's a very fundamental thing.”
November 25, 2025 at 8:58 PM
Straight from the horse's mouth:
“These models somehow just generalize dramatically worse than people. It's a very fundamental thing.”
“These models somehow just generalize dramatically worse than people. It's a very fundamental thing.”
Reposted by Michael
Ironwood, the 7th generation of Google's TPUs cloud.google.com/blog/product...
November 26, 2025 at 6:49 AM
Ironwood, the 7th generation of Google's TPUs cloud.google.com/blog/product...
The AI Otter, he’s a playful one. They are capable of searching the depths and surfacing wonders, but utility is hard to extract from them.
November 21, 2025 at 1:57 AM
The AI Otter, he’s a playful one. They are capable of searching the depths and surfacing wonders, but utility is hard to extract from them.
Reposted by Michael
Literally the subtitle to Mary Shelley’s “Frankenstein.”
November 17, 2025 at 1:19 PM
Literally the subtitle to Mary Shelley’s “Frankenstein.”
Reposted by Michael
Jeff Bezos will be co-CEO of AI startup Project Prometheus
Jeff Bezos will be co-CEO of AI startup Project Prometheus
Jeff Bezos has a new job.
buff.ly
November 17, 2025 at 10:20 AM
Jeff Bezos will be co-CEO of AI startup Project Prometheus
Reposted by Michael
Stateful Agents Meetup: Learn how Letta Code was built.
Thursday, November 20th in San Francisco.
We'll do a technical deep dive on building persistence and long-term learning into CLI agents. Meet the team, see how it works under the hood.
luma.com/stateful-ag...
Thursday, November 20th in San Francisco.
We'll do a technical deep dive on building persistence and long-term learning into CLI agents. Meet the team, see how it works under the hood.
luma.com/stateful-ag...
Stateful Agents Meetup: Building Stateful CLI Agents · Luma
Building Stateful CLI Agents
CLI agents like Claude Code, Gemini CLI, and Codex are perhaps the closest we've gotten to real, autonomous agents that can…
luma.com
November 18, 2025 at 12:52 AM
Stateful Agents Meetup: Learn how Letta Code was built.
Thursday, November 20th in San Francisco.
We'll do a technical deep dive on building persistence and long-term learning into CLI agents. Meet the team, see how it works under the hood.
luma.com/stateful-ag...
Thursday, November 20th in San Francisco.
We'll do a technical deep dive on building persistence and long-term learning into CLI agents. Meet the team, see how it works under the hood.
luma.com/stateful-ag...