Sam Harsimony
@harsimony.bsky.social
I write about opportunities in science, space, and policy here: https://splittinginfinity.substack.com/
Pinned
Sam Harsimony
@harsimony.bsky.social
· Jan 26
Thread of my new posts in the replies (and some summaries of old ones).
Reposted by Sam Harsimony
With Devstral 2, Mistral pushes the envelope on two fronts:
① what an agentic coding model can do, with no reasoning!
There’s a quality gap to reasoning models, expectedly. The positive: it is cheaper; potentially even cheaper in practice than indicated in this chart.
① what an agentic coding model can do, with no reasoning!
There’s a quality gap to reasoning models, expectedly. The positive: it is cheaper; potentially even cheaper in practice than indicated in this chart.
December 9, 2025 at 5:21 PM
With Devstral 2, Mistral pushes the envelope on two fronts:
① what an agentic coding model can do, with no reasoning!
There’s a quality gap to reasoning models, expectedly. The positive: it is cheaper; potentially even cheaper in practice than indicated in this chart.
① what an agentic coding model can do, with no reasoning!
There’s a quality gap to reasoning models, expectedly. The positive: it is cheaper; potentially even cheaper in practice than indicated in this chart.
Apparently Scott Aaronson figured out how to watermark LLM outputs so that inference providers can check if a piece of text came from their LLM while being undetectable to others.
But OpenAI didn't deploy it. Google does something like this though.
scottaaronson.blog?p=9333
But OpenAI didn't deploy it. Google does something like this though.
scottaaronson.blog?p=9333
Theory and AI Alignment
The following is based on a talk that I gave (remotely) at the UK AI Safety Institute Alignment Workshop on October 29, and which I then procrastinated for more than a month in writing up. Enjoy! T…
scottaaronson.blog
December 9, 2025 at 10:38 PM
Apparently Scott Aaronson figured out how to watermark LLM outputs so that inference providers can check if a piece of text came from their LLM while being undetectable to others.
But OpenAI didn't deploy it. Google does something like this though.
scottaaronson.blog?p=9333
But OpenAI didn't deploy it. Google does something like this though.
scottaaronson.blog?p=9333
"In 2023, Nigeria had a million more births than *the whole of Europe*."
Humanity's future is quite literally in Sub-Saharan Africa.
(1/3)
Humanity's future is quite literally in Sub-Saharan Africa.
(1/3)
December 9, 2025 at 12:28 AM
"In 2023, Nigeria had a million more births than *the whole of Europe*."
Humanity's future is quite literally in Sub-Saharan Africa.
(1/3)
Humanity's future is quite literally in Sub-Saharan Africa.
(1/3)
Reposted by Sam Harsimony
Here's a great review of what we saw in AI this year, from @gleech.org
AI in 2025: gestalt — LessWrong
This is the editorial for this year’s "Shallow Review of AI Safety". (It got long enough to stand alone.) …
www.lesswrong.com
December 8, 2025 at 5:24 PM
Here's a great review of what we saw in AI this year, from @gleech.org
Reposted by Sam Harsimony
It was never conceived this way, but most of the idealized consequences of "solarpunk" can be a direct consequence of this line going down. Radically decentralized greened desert communities can be delivered directly by market forces. caseyhandmer.wordpress.com/2025/12/08/e...
Energy Predictions 2025
Printable pdf. It’s been a few years since I wrote a broad post on energy, so I’m providing an update in one easy to read place. More detailed specific posts on energy are here. If you want t…
caseyhandmer.wordpress.com
December 8, 2025 at 10:25 PM
It was never conceived this way, but most of the idealized consequences of "solarpunk" can be a direct consequence of this line going down. Radically decentralized greened desert communities can be delivered directly by market forces. caseyhandmer.wordpress.com/2025/12/08/e...
Reposted by Sam Harsimony
December 5, 2025 at 2:43 PM
Reposted by Sam Harsimony
In 2023, papers denying motherhood’s role in the gender wage gap went viral. I called out their flawed assumptions then, and newer results have proven me right.
www.writingruxandrabio.com/p/of-course-...
www.writingruxandrabio.com/p/of-course-...
Of course motherhood drives the gender wage gap
Discussions around the motherhood wage gap reveal the limits of autism and the benefits of female intuition
www.writingruxandrabio.com
December 6, 2025 at 10:11 PM
In 2023, papers denying motherhood’s role in the gender wage gap went viral. I called out their flawed assumptions then, and newer results have proven me right.
www.writingruxandrabio.com/p/of-course-...
www.writingruxandrabio.com/p/of-course-...
It would be neat to see the feed below a post branch into two timelines: one where you liked the post, and one where you didn't.
Makes it very obvious what kind of cage you're building for yourself.
Makes it very obvious what kind of cage you're building for yourself.
The hyper-responsiveness of the For You feed is training me not to click on ragebait. It’s like a little automated Yoda designed to prove that anger leads to hate leads to suffering.
December 4, 2025 at 10:20 PM
It would be neat to see the feed below a post branch into two timelines: one where you liked the post, and one where you didn't.
Makes it very obvious what kind of cage you're building for yourself.
Makes it very obvious what kind of cage you're building for yourself.
Reposted by Sam Harsimony
New post: The Eldest Millennials Had the Same Fertility as the Youngest Baby Boomers.
I did not have a strong take digging into 'fertility crisis' debates, but was genuinely surprised as how different the data is from the panic of that Discourse.
mikekonczal.substack.com/p/the-eldest...
I did not have a strong take digging into 'fertility crisis' debates, but was genuinely surprised as how different the data is from the panic of that Discourse.
mikekonczal.substack.com/p/the-eldest...
The Eldest Millennials Had the Same Fertility as the Youngest Baby Boomers
How U.S. fertility is happening later, not less.
mikekonczal.substack.com
December 4, 2025 at 1:12 PM
New post: The Eldest Millennials Had the Same Fertility as the Youngest Baby Boomers.
I did not have a strong take digging into 'fertility crisis' debates, but was genuinely surprised as how different the data is from the panic of that Discourse.
mikekonczal.substack.com/p/the-eldest...
I did not have a strong take digging into 'fertility crisis' debates, but was genuinely surprised as how different the data is from the panic of that Discourse.
mikekonczal.substack.com/p/the-eldest...
Reposted by Sam Harsimony
on balance, i think i prefer such power to be equalized and open-sourced than held by a few API providers even if that means opening pandora's box
so whether or not you like AI, i think we need to realize how scary and dangerous things like this new zimage model are. seriously.
December 4, 2025 at 5:33 AM
on balance, i think i prefer such power to be equalized and open-sourced than held by a few API providers even if that means opening pandora's box
This is exactly what you'd do if you were trying to sell everyone a laptop with AI on it.
With 128x compression English Wikipedia fits in 0.2 GB.
With 128x compression English Wikipedia fits in 0.2 GB.
Apple's CLaRa-7B-Instruct (Compression-16 & 128)
The CLaRa-7B-Instruct model is Apple's instruction-tuned unified RAG model with built-in semantic document compression (16× & 128x). It supports instruction-following QA directly from compressed document representations.
The CLaRa-7B-Instruct model is Apple's instruction-tuned unified RAG model with built-in semantic document compression (16× & 128x). It supports instruction-following QA directly from compressed document representations.
December 3, 2025 at 9:37 PM
This is exactly what you'd do if you were trying to sell everyone a laptop with AI on it.
With 128x compression English Wikipedia fits in 0.2 GB.
With 128x compression English Wikipedia fits in 0.2 GB.
Reposted by Sam Harsimony
My AI Safety Paper Highlights of November 2025:
- *Natural emergent misalignment*
- Honesty interventions, lie detection
- Self-report finetuning
- CoT obfuscation from output monitors
- Consistency training for robustness
- Weight-space steering
More at open.substack.com/pub/aisafety...
- *Natural emergent misalignment*
- Honesty interventions, lie detection
- Self-report finetuning
- CoT obfuscation from output monitors
- Consistency training for robustness
- Weight-space steering
More at open.substack.com/pub/aisafety...
Paper Highlights of November 2025
Natural emergent misalignment, honesty interventions, self-report finetuning, CoT obfuscation from output monitors, consistency training for robustness, and weight-space steering
open.substack.com
December 2, 2025 at 9:05 PM
My AI Safety Paper Highlights of November 2025:
- *Natural emergent misalignment*
- Honesty interventions, lie detection
- Self-report finetuning
- CoT obfuscation from output monitors
- Consistency training for robustness
- Weight-space steering
More at open.substack.com/pub/aisafety...
- *Natural emergent misalignment*
- Honesty interventions, lie detection
- Self-report finetuning
- CoT obfuscation from output monitors
- Consistency training for robustness
- Weight-space steering
More at open.substack.com/pub/aisafety...
Reposted by Sam Harsimony
Wow.
"It gives me no pleasure to say what I’m about to say because I worked with Pete Hegseth for seven or eight years at Fox News. This is an act of a war crime .... There’s absolutely no legal basis for it.”
- Newsmax's Judge Napolitano
"It gives me no pleasure to say what I’m about to say because I worked with Pete Hegseth for seven or eight years at Fox News. This is an act of a war crime .... There’s absolutely no legal basis for it.”
- Newsmax's Judge Napolitano
Woah. Newsmax’s legal analyst just said Pete Hegseth and everyone involved in the illegal boat strike should be “prosecuted for a war crime.”
They’ve even lost Newsmax on this one.
They’ve even lost Newsmax on this one.
December 3, 2025 at 2:06 AM
Wow.
"It gives me no pleasure to say what I’m about to say because I worked with Pete Hegseth for seven or eight years at Fox News. This is an act of a war crime .... There’s absolutely no legal basis for it.”
- Newsmax's Judge Napolitano
"It gives me no pleasure to say what I’m about to say because I worked with Pete Hegseth for seven or eight years at Fox News. This is an act of a war crime .... There’s absolutely no legal basis for it.”
- Newsmax's Judge Napolitano
More on the Pivotal eVTOL. Main barrier for use in cities new seems to be noise. At altitude it's not bad, but sounds like weed wackers at landing.
Larger propellers with optimized shape and soundproofed landing areas might work.
www.youtube.com/watch?v=oT80...
Larger propellers with optimized shape and soundproofed landing areas might work.
www.youtube.com/watch?v=oT80...
Going UP!!! Flying A VTOL For Real - Pivotal's Unique Take on Accessible Aviation
YouTube video by Scott Manley
www.youtube.com
December 2, 2025 at 8:04 PM
More on the Pivotal eVTOL. Main barrier for use in cities new seems to be noise. At altitude it's not bad, but sounds like weed wackers at landing.
Larger propellers with optimized shape and soundproofed landing areas might work.
www.youtube.com/watch?v=oT80...
Larger propellers with optimized shape and soundproofed landing areas might work.
www.youtube.com/watch?v=oT80...
You can run a 7B model on your laptop. Leading open-weight models are ~700B params.
People's hardware will get better and models of constant capability will shrink. Perhaps every laptop will run models with capability equal to today's frontier.
People's hardware will get better and models of constant capability will shrink. Perhaps every laptop will run models with capability equal to today's frontier.
December 2, 2025 at 6:25 PM
You can run a 7B model on your laptop. Leading open-weight models are ~700B params.
People's hardware will get better and models of constant capability will shrink. Perhaps every laptop will run models with capability equal to today's frontier.
People's hardware will get better and models of constant capability will shrink. Perhaps every laptop will run models with capability equal to today's frontier.
Reposted by Sam Harsimony
“I will die on the hill that population coding is the relevant level of encoding information in the brain.” In the latest “This paper changed my life,” Nancy Padilla-Coreano discusses a paper on mixed selectivity neurons.
#neuroskyence
www.thetransmitter.org/this-paper-c...
#neuroskyence
www.thetransmitter.org/this-paper-c...
This paper changed my life: Nancy Padilla-Coreano on learning the value of population coding
The 2013 Nature paper by Mattia Rigotti and his colleagues revealed how mixed selectivity neurons—cells that are not selectively tuned to a stimulus—play a key role in cognition.
www.thetransmitter.org
December 1, 2025 at 2:23 PM
“I will die on the hill that population coding is the relevant level of encoding information in the brain.” In the latest “This paper changed my life,” Nancy Padilla-Coreano discusses a paper on mixed selectivity neurons.
#neuroskyence
www.thetransmitter.org/this-paper-c...
#neuroskyence
www.thetransmitter.org/this-paper-c...
Reposted by Sam Harsimony
DeepSeek released V3.2 (and V3.2 Speciale, a math-oriented model).
New model, new benchmarks!
The biggest jump for DeepSeek V3.2 is on agentic coding, where it seems poised to erase a lot of models on the Pareto frontier, including Sonnet 4.5, Minimax M2, and K2 Thinking.
New model, new benchmarks!
The biggest jump for DeepSeek V3.2 is on agentic coding, where it seems poised to erase a lot of models on the Pareto frontier, including Sonnet 4.5, Minimax M2, and K2 Thinking.
December 1, 2025 at 6:28 PM
DeepSeek released V3.2 (and V3.2 Speciale, a math-oriented model).
New model, new benchmarks!
The biggest jump for DeepSeek V3.2 is on agentic coding, where it seems poised to erase a lot of models on the Pareto frontier, including Sonnet 4.5, Minimax M2, and K2 Thinking.
New model, new benchmarks!
The biggest jump for DeepSeek V3.2 is on agentic coding, where it seems poised to erase a lot of models on the Pareto frontier, including Sonnet 4.5, Minimax M2, and K2 Thinking.
Brain emulation is quietly getting better. Important to keep an eye on.
arxiv.org/abs/2510.15745
arxiv.org/abs/2510.15745
State of Brain Emulation Report 2025
The State of Brain Emulation Report 2025 provides a comprehensive reassessment of the field's progress since Sandberg and Bostrom's 2008 Whole Brain Emulation roadmap. The report is organized around t...
arxiv.org
November 30, 2025 at 4:25 PM
Brain emulation is quietly getting better. Important to keep an eye on.
arxiv.org/abs/2510.15745
arxiv.org/abs/2510.15745
Reposted by Sam Harsimony
Just when I put out a linkpost Tensor Economics blog puts up another Banger.
Appears to be a more rigorous look at RL-as-a-Service, among other things
www.tensoreconomics.com/p/ai-infrast...
Appears to be a more rigorous look at RL-as-a-Service, among other things
www.tensoreconomics.com/p/ai-infrast...
AI infrastructure in the "Era of experience"
Intelligence involution, economies of scale in RL, everything async and multi-turn.
www.tensoreconomics.com
November 26, 2025 at 9:44 PM
Just when I put out a linkpost Tensor Economics blog puts up another Banger.
Appears to be a more rigorous look at RL-as-a-Service, among other things
www.tensoreconomics.com/p/ai-infrast...
Appears to be a more rigorous look at RL-as-a-Service, among other things
www.tensoreconomics.com/p/ai-infrast...
Evidence for my view that next phase of AI is drilling down on different domains. This is The Way.
bsky.app/profile/hars...
bsky.app/profile/hars...
Gemini 3 Pro set a new record on GPQA Diamond: 93% vs. the previous record of 88%. What you can’t tell from the headline: almost all of this gain came in organic chemistry. 🧬🧵
November 26, 2025 at 3:46 PM
Evidence for my view that next phase of AI is drilling down on different domains. This is The Way.
bsky.app/profile/hars...
bsky.app/profile/hars...