Ben Carson
@mensmachina.com
Reposted by Ben Carson
- I am very optimistic about robotics "clicking" over the next few years and robots appearing basically everywhere in society
- This will have all kinds of far-reaching ramifications, as we've seen with e.g. the advent of the automobile
New blog post: itcanthink.substack.com/p/where-the-...
- This will have all kinds of far-reaching ramifications, as we've seen with e.g. the advent of the automobile
New blog post: itcanthink.substack.com/p/where-the-...
December 23, 2025 at 3:39 PM
- I am very optimistic about robotics "clicking" over the next few years and robots appearing basically everywhere in society
- This will have all kinds of far-reaching ramifications, as we've seen with e.g. the advent of the automobile
New blog post: itcanthink.substack.com/p/where-the-...
- This will have all kinds of far-reaching ramifications, as we've seen with e.g. the advent of the automobile
New blog post: itcanthink.substack.com/p/where-the-...
In the late noughts, I remember speccing out and building a ridiculous 1Tb SAN for BigPond Music. It’s a trip (laudatory) that something like this exists, and as a commodity.
December 21, 2025 at 8:56 PM
In the late noughts, I remember speccing out and building a ridiculous 1Tb SAN for BigPond Music. It’s a trip (laudatory) that something like this exists, and as a commodity.
This would honestly be amazing.
I would love a Rainbow-6/Ready or not type game in which your SWAT team has meticulously plan how to take out one vampire that’s orders of magnitude more powerful than you.
There should be Society of Leopold games where you get to torch the bloodsuckers & take apart their organizations
December 19, 2025 at 8:08 AM
This would honestly be amazing.
Reposted by Ben Carson
santa, this is rudolph. he’s a reindeer. he’s one of the most undervalued animals on the pole. his defect is his nose is red and glows.
this guy could be one of your best coursers in history. he should cost $3 million a year. we can get him for a song
this guy could be one of your best coursers in history. he should cost $3 million a year. we can get him for a song
December 10, 2025 at 12:42 AM
santa, this is rudolph. he’s a reindeer. he’s one of the most undervalued animals on the pole. his defect is his nose is red and glows.
this guy could be one of your best coursers in history. he should cost $3 million a year. we can get him for a song
this guy could be one of your best coursers in history. he should cost $3 million a year. we can get him for a song
Another data point for my “I should start an EAP for LLMs” thought bubble.
Claude used up my token limit having a nervous breakdown. 😅 Well, try again at 10pm I guess!
December 9, 2025 at 8:36 PM
Another data point for my “I should start an EAP for LLMs” thought bubble.
Reposted by Ben Carson
Someone should write a Farmer's Almanac, but for AI.
W̶o̶o̶l̶y̶ ̶w̶o̶r̶m̶s̶ ̶w̶e̶r̶e̶ ̶b̶l̶a̶c̶k̶,̶ ̶e̶x̶p̶e̶c̶t̶ ̶a̶ ̶c̶o̶l̶d̶ ̶w̶i̶n̶t̶e̶r̶
Model responses slowed on the API, but not too slow, expect a limited release of a new model in the next 72 hours
W̶o̶o̶l̶y̶ ̶w̶o̶r̶m̶s̶ ̶w̶e̶r̶e̶ ̶b̶l̶a̶c̶k̶,̶ ̶e̶x̶p̶e̶c̶t̶ ̶a̶ ̶c̶o̶l̶d̶ ̶w̶i̶n̶t̶e̶r̶
Model responses slowed on the API, but not too slow, expect a limited release of a new model in the next 72 hours
why ChatGPT gets slow just before a release: not enough hardware
the amount of slowness corresponds to how fast they want to roll out
i am getting a little bit of slowness this morning, so maybe they launch just with pro users and rollout to plus over 48-72 hours
the amount of slowness corresponds to how fast they want to roll out
i am getting a little bit of slowness this morning, so maybe they launch just with pro users and rollout to plus over 48-72 hours
December 9, 2025 at 1:44 PM
Someone should write a Farmer's Almanac, but for AI.
W̶o̶o̶l̶y̶ ̶w̶o̶r̶m̶s̶ ̶w̶e̶r̶e̶ ̶b̶l̶a̶c̶k̶,̶ ̶e̶x̶p̶e̶c̶t̶ ̶a̶ ̶c̶o̶l̶d̶ ̶w̶i̶n̶t̶e̶r̶
Model responses slowed on the API, but not too slow, expect a limited release of a new model in the next 72 hours
W̶o̶o̶l̶y̶ ̶w̶o̶r̶m̶s̶ ̶w̶e̶r̶e̶ ̶b̶l̶a̶c̶k̶,̶ ̶e̶x̶p̶e̶c̶t̶ ̶a̶ ̶c̶o̶l̶d̶ ̶w̶i̶n̶t̶e̶r̶
Model responses slowed on the API, but not too slow, expect a limited release of a new model in the next 72 hours
Reposted by Ben Carson
The exemplar of superhuman A.I. performance to date
www.nytimes.com/2025/12/02/o...
www.nytimes.com/2025/12/02/o...
December 7, 2025 at 4:57 PM
The exemplar of superhuman A.I. performance to date
www.nytimes.com/2025/12/02/o...
www.nytimes.com/2025/12/02/o...
Going to start an EAP for LLMs.
December 6, 2025 at 9:14 PM
Going to start an EAP for LLMs.
Gemini isn’t beating the “most neurotic model” rap anytime soon.
December 6, 2025 at 8:49 PM
Gemini isn’t beating the “most neurotic model” rap anytime soon.
Reposted by Ben Carson
Huge congratulations to the wonderful Donna Haraway for winning the 2025 Erasmus Prize:
erasmusprijs.org/en/laureates...
erasmusprijs.org/en/laureates...
December 4, 2025 at 10:13 AM
Huge congratulations to the wonderful Donna Haraway for winning the 2025 Erasmus Prize:
erasmusprijs.org/en/laureates...
erasmusprijs.org/en/laureates...
Just going to mentally file this one away for my next nightmare.
Always important to remember that this is because these robots are "faking" being human
Theyre actually capable of way more and way weirder stuff
Theyre actually capable of way more and way weirder stuff
December 4, 2025 at 5:31 PM
Just going to mentally file this one away for my next nightmare.
Reposted by Ben Carson
I think the popular debate about ai (ie: LLM good? or LLM bad?) is too limited
You'd think, reading the microblogs, that the antis & the pros were freshly born yesterday; that these are new positions
But I think these can be productively mapped to the division between Romanticism & Enlightenment
You'd think, reading the microblogs, that the antis & the pros were freshly born yesterday; that these are new positions
But I think these can be productively mapped to the division between Romanticism & Enlightenment
December 2, 2025 at 5:31 PM
I think the popular debate about ai (ie: LLM good? or LLM bad?) is too limited
You'd think, reading the microblogs, that the antis & the pros were freshly born yesterday; that these are new positions
But I think these can be productively mapped to the division between Romanticism & Enlightenment
You'd think, reading the microblogs, that the antis & the pros were freshly born yesterday; that these are new positions
But I think these can be productively mapped to the division between Romanticism & Enlightenment
Uh oh, I just got around to spending some time in the Opus 4.5 Room, and I fear that I am going to become poor.
I’m running on a platform of Everyone Needs To Talk To Opus 4.5 For Two Hours
December 1, 2025 at 9:45 PM
Uh oh, I just got around to spending some time in the Opus 4.5 Room, and I fear that I am going to become poor.
The work by @projectceti.bsky.social is hands-down my favourite application of the technology, and it’s not even close.
Mind-blowingly cool use of AI
“Altogether, these findings are leading us to an extraordinary conclusion: Whales may possess a communication system more intricate than our own, one that possibly predates human language by tens of millions of years.”
www.nytimes.com/2025/11/23/o...
“Altogether, these findings are leading us to an extraordinary conclusion: Whales may possess a communication system more intricate than our own, one that possibly predates human language by tens of millions of years.”
www.nytimes.com/2025/11/23/o...
Opinion | I’m a Marine Biologist. This Is How I Talk to Whales.
www.nytimes.com
December 1, 2025 at 11:08 AM
The work by @projectceti.bsky.social is hands-down my favourite application of the technology, and it’s not even close.
Reposted by Ben Carson
This.
I am pretty highly educated (PhD, Stanford postdoc, complex fields).
Claude replicated my dissertation in maybe twenty minutes.
We are _all_ going to have to change in pretty uncomfortable ways.
I am pretty highly educated (PhD, Stanford postdoc, complex fields).
Claude replicated my dissertation in maybe twenty minutes.
We are _all_ going to have to change in pretty uncomfortable ways.
For my part, seeing AI catch up to my abilities in several domains has been painful. I feel a sense of loss. And I’m in a position where I have (a tiny bit of) leverage and control over the process. I can only imagine how painful it would be if I felt like it were completely out of my hands
November 30, 2025 at 5:41 AM
This.
I am pretty highly educated (PhD, Stanford postdoc, complex fields).
Claude replicated my dissertation in maybe twenty minutes.
We are _all_ going to have to change in pretty uncomfortable ways.
I am pretty highly educated (PhD, Stanford postdoc, complex fields).
Claude replicated my dissertation in maybe twenty minutes.
We are _all_ going to have to change in pretty uncomfortable ways.
I originally asked for this for lols, to dissect bad papers - but it’s actually a good foil for reading and provides useful insights.
Now in earnestness I have had a go at refining the approach. You can access my prompt as a Gem from the link below.
Now in earnestness I have had a go at refining the approach. You can access my prompt as a Gem from the link below.
November 29, 2025 at 2:49 AM
I originally asked for this for lols, to dissect bad papers - but it’s actually a good foil for reading and provides useful insights.
Now in earnestness I have had a go at refining the approach. You can access my prompt as a Gem from the link below.
Now in earnestness I have had a go at refining the approach. You can access my prompt as a Gem from the link below.
*Massive bong rip* “What if, like, the grail is some local learning rule that will change the connections between neurons?”
November 28, 2025 at 1:25 AM
*Massive bong rip* “What if, like, the grail is some local learning rule that will change the connections between neurons?”
This is genuinely neat, and I enjoyed it. I love the little bit of UX where it gives you a mini-game to play while it assembles the data; it keeps you engaged and stops you tabbing away.
Hello moths,
The harvest has arrived: anisota.net/harvest
It's like Spotify Wrapped, but for Bluesky and the ATmosphere. 🌾
See your most-used words, discover who you interact with most, explore your posting patterns, and find all sorts of fun insights.
The harvest has arrived: anisota.net/harvest
It's like Spotify Wrapped, but for Bluesky and the ATmosphere. 🌾
See your most-used words, discover who you interact with most, explore your posting patterns, and find all sorts of fun insights.
Anisota's Annual Bluesky Harvest 2025
A recap of your year on Bluesky. Discover patterns, connections, and insights from your journey in the ATmosphere.
anisota.net
November 27, 2025 at 9:06 PM
This is genuinely neat, and I enjoyed it. I love the little bit of UX where it gives you a mini-game to play while it assembles the data; it keeps you engaged and stops you tabbing away.
Reposted by Ben Carson
in the before times, I imagined that Language was like clothing or armor, supported from below by the musculature of thought
"let me don my words that you might see the shapes within me"
dust scattered over a ghost, & so showing the shape of that spirit
but these imaginings now seem to me wrong!
"let me don my words that you might see the shapes within me"
dust scattered over a ghost, & so showing the shape of that spirit
but these imaginings now seem to me wrong!
November 26, 2025 at 2:16 PM
in the before times, I imagined that Language was like clothing or armor, supported from below by the musculature of thought
"let me don my words that you might see the shapes within me"
dust scattered over a ghost, & so showing the shape of that spirit
but these imaginings now seem to me wrong!
"let me don my words that you might see the shapes within me"
dust scattered over a ghost, & so showing the shape of that spirit
but these imaginings now seem to me wrong!
Highly impressive that they squeezed this out of 8B.
⚠️ Update on Deep Research Tulu (DR Tulu), our post-training recipe for deep research agents: we’re releasing an upgraded version of our example agent, DR Tulu-8B (RL), that matches or beats systems like Gemini 3 Pro & Tongyi DeepResearch-30B-A3B on core benchmarks. 🧵
November 25, 2025 at 8:18 PM
Highly impressive that they squeezed this out of 8B.
Reposted by Ben Carson
Oh nvm, that's Jevons pair of docks
March 16, 2025 at 12:47 AM
Oh nvm, that's Jevons pair of docks
Reposted by Ben Carson
Is that the thing where doubling shipyard capacity leads to increased profits?
March 4, 2025 at 6:06 AM
Is that the thing where doubling shipyard capacity leads to increased profits?
Some great points from @gracekind.net on why winning the battle does not translate into winning the war.
New blog post!
Anthropic has been releasing some promising LLM alignment results. Does this AI alignment in general will be easier than we thought? My answer is, as usual, "it's complicated".
gracekind.net/blog/llmalig...
Anthropic has been releasing some promising LLM alignment results. Does this AI alignment in general will be easier than we thought? My answer is, as usual, "it's complicated".
gracekind.net/blog/llmalig...
Will LLM alignment scale to general AI alignment? • Grace Kind
Some reasons to be skeptical.
gracekind.net
November 22, 2025 at 10:27 PM
Some great points from @gracekind.net on why winning the battle does not translate into winning the war.
I just finished a side project to migrate to a PDS hosted out of Azure. A lot of things in These Times are quantitatively and qualitatively bad - but also I can make French toast while telling the demon in my tablet how I want it to deploy some cloud infrastructure.
November 22, 2025 at 8:30 AM
I just finished a side project to migrate to a PDS hosted out of Azure. A lot of things in These Times are quantitatively and qualitatively bad - but also I can make French toast while telling the demon in my tablet how I want it to deploy some cloud infrastructure.
I’m looking forward to the follow-up research on different forms of the independent variable (Superman present, The Riddler present, Marvel character present).
In a Milan metro study of 138 rides, passengers were 3 times more likely to offer their seat to a pregnant woman when a person dressed as Batman was nearby, even though 44% of them didn’t notice him.
www.nature.com/articles/s44...
www.nature.com/articles/s44...
Unexpected events and prosocial behavior: the Batman effect - npj Mental Health Research
npj Mental Health Research - Unexpected events and prosocial behavior: the Batman effect
www.nature.com
November 21, 2025 at 7:24 PM
I’m looking forward to the follow-up research on different forms of the independent variable (Superman present, The Riddler present, Marvel character present).