Martin Singh-Blom
@singhblom.com
Head of ML at Embark Studios. Working on RL for physics based animation, LLM applications in games, generative models, etc.
I like it when ML makes what previously was impossible possible. Previously web dev, genetics, mathematics, physics.
I like it when ML makes what previously was impossible possible. Previously web dev, genetics, mathematics, physics.
Reposted by Martin Singh-Blom
Incredible work by the Dallas Fed.
October 11, 2025 at 5:36 PM
Incredible work by the Dallas Fed.
Reposted by Martin Singh-Blom
Soooooo when is bluesky deploying bridging systems
arxiv.org/abs/2301.09976
arxiv.org/abs/2301.09976
Bridging Systems: Open Problems for Countering Destructive Divisiveness across Ranking, Recommenders, and Governance
Divisiveness appears to be increasing in much of the world, leading to concern about political violence and a decreasing capacity to collaboratively address large-scale societal challenges. In this wo...
arxiv.org
October 8, 2025 at 4:15 AM
Soooooo when is bluesky deploying bridging systems
arxiv.org/abs/2301.09976
arxiv.org/abs/2301.09976
Reposted by Martin Singh-Blom
I don't really polisci post much anymore because it's like "Hi, political scientist here. I'm a big brain expert who ran 5 trillion lines of code in R and read 900 books and I'm here to tell you it's bad when masked stormtroopers kidnap your baby. Don't forget to like and follow."
October 5, 2025 at 2:42 PM
I don't really polisci post much anymore because it's like "Hi, political scientist here. I'm a big brain expert who ran 5 trillion lines of code in R and read 900 books and I'm here to tell you it's bad when masked stormtroopers kidnap your baby. Don't forget to like and follow."
Reposted by Martin Singh-Blom
I disagree with many moderation decisions from the Bluesky team, but at the same time the toxicity towards the team is making me pretty uncomfortable.
I think the priority should be for them to fix this 👇🏼 and for us to start using ATProto as intended and own our own infra and moderation.
I think the priority should be for them to fix this 👇🏼 and for us to start using ATProto as intended and own our own infra and moderation.
October 3, 2025 at 3:59 PM
I disagree with many moderation decisions from the Bluesky team, but at the same time the toxicity towards the team is making me pretty uncomfortable.
I think the priority should be for them to fix this 👇🏼 and for us to start using ATProto as intended and own our own infra and moderation.
I think the priority should be for them to fix this 👇🏼 and for us to start using ATProto as intended and own our own infra and moderation.
Reposted by Martin Singh-Blom
🇺🇦🇫🇷 Zelensky: The distance between Moscow and Paris is less than 3,000 km, and the missiles that Russia uses to shell Ukraine have a range of 2,500 km.
In two years the Russians will have many missiles with a range of 5,000 km. The sea will protect no one, the ocean will protect no one.
In two years the Russians will have many missiles with a range of 5,000 km. The sea will protect no one, the ocean will protect no one.
September 4, 2025 at 6:36 AM
🇺🇦🇫🇷 Zelensky: The distance between Moscow and Paris is less than 3,000 km, and the missiles that Russia uses to shell Ukraine have a range of 2,500 km.
In two years the Russians will have many missiles with a range of 5,000 km. The sea will protect no one, the ocean will protect no one.
In two years the Russians will have many missiles with a range of 5,000 km. The sea will protect no one, the ocean will protect no one.
Reposted by Martin Singh-Blom
Longcat-Flash-Chat (560B)
uh, holy shit this one is intriguing. bare minimum they compare themselves to all the (actual) top models and do okay
but inside.. damn this one has some cool ideas
huggingface.co/meituan-long...
uh, holy shit this one is intriguing. bare minimum they compare themselves to all the (actual) top models and do okay
but inside.. damn this one has some cool ideas
huggingface.co/meituan-long...
August 31, 2025 at 11:20 AM
Longcat-Flash-Chat (560B)
uh, holy shit this one is intriguing. bare minimum they compare themselves to all the (actual) top models and do okay
but inside.. damn this one has some cool ideas
huggingface.co/meituan-long...
uh, holy shit this one is intriguing. bare minimum they compare themselves to all the (actual) top models and do okay
but inside.. damn this one has some cool ideas
huggingface.co/meituan-long...
Reposted by Martin Singh-Blom
After we've read 'In 80 days around the world' my 8y old just remarked that if Phileas Fogg had taken the Deutsche Bahn, he would have needed some extra buffer time 😅
DB, you have a serious reputation issue with the young generation...
DB, you have a serious reputation issue with the young generation...
August 6, 2025 at 6:31 PM
After we've read 'In 80 days around the world' my 8y old just remarked that if Phileas Fogg had taken the Deutsche Bahn, he would have needed some extra buffer time 😅
DB, you have a serious reputation issue with the young generation...
DB, you have a serious reputation issue with the young generation...
Reposted by Martin Singh-Blom
I don’t know who needs this right now, but I really enjoy browsing “Paper Skygest”, a feed of posts on academic papers *from people you follow*. I think it’s what a lot of us are actually here for
August 6, 2025 at 11:35 AM
I don’t know who needs this right now, but I really enjoy browsing “Paper Skygest”, a feed of posts on academic papers *from people you follow*. I think it’s what a lot of us are actually here for
I completely agree with this. I don't like the man's politics at all but I have no doubt that it's _his_ politics we'll get regardless of if he asks ChatGPT about stuff or not.
The Swedish prime minister has faced criticism for saying he gets advice from ChatGPT about running the country.
I think this outrage is overblown as using ChatGPT for feedback and advice is already a normal part of many people’s lives which will only grow. That genie isn’t going back in the bottle
I think this outrage is overblown as using ChatGPT for feedback and advice is already a normal part of many people’s lives which will only grow. That genie isn’t going back in the bottle
‘We didn’t vote for ChatGPT’: Swedish PM under fire for using AI in role
Tech experts criticise Ulf Kristersson as newspaper accuses him of falling for ‘the oligarchs’ AI psychosis’
www.theguardian.com
August 6, 2025 at 9:30 AM
I completely agree with this. I don't like the man's politics at all but I have no doubt that it's _his_ politics we'll get regardless of if he asks ChatGPT about stuff or not.
Reposted by Martin Singh-Blom
Reposted by Martin Singh-Blom
Every time someone tells me to use fewer em dashes ⸻ I bring out Big Bertha.
That's how you can tell human authors from AI.
AI will put in some em dashes, then apologise and take them out when told off.
Human authors will climb the filing cabinet, hissing, and defend their right to em dashes with a loaded stapler.
AI will put in some em dashes, then apologise and take them out when told off.
Human authors will climb the filing cabinet, hissing, and defend their right to em dashes with a loaded stapler.
Em dashes, like anything, should be used in moderation. Check first with your editor if em dashes are right for you. 😉
#WriteSky #WritingTips
#WriteSky #WritingTips
August 3, 2025 at 2:19 AM
Every time someone tells me to use fewer em dashes ⸻ I bring out Big Bertha.
Reposted by Martin Singh-Blom
We've seen this story a lot. Personal computers were niche in the 80s. Hard to use, it seemed obvs that they would remain toys for nerds.
Cell phones were niche, and for the rich.
I think the downside prediction for widespread on-device AI is "it takes a decade." "Can't happen" is not an outcome.
Cell phones were niche, and for the rich.
I think the downside prediction for widespread on-device AI is "it takes a decade." "Can't happen" is not an outcome.
August 2, 2025 at 7:01 PM
We've seen this story a lot. Personal computers were niche in the 80s. Hard to use, it seemed obvs that they would remain toys for nerds.
Cell phones were niche, and for the rich.
I think the downside prediction for widespread on-device AI is "it takes a decade." "Can't happen" is not an outcome.
Cell phones were niche, and for the rich.
I think the downside prediction for widespread on-device AI is "it takes a decade." "Can't happen" is not an outcome.
Reposted by Martin Singh-Blom
Bro, just one more humanoid robot, bro please, this one will be the true general purpose technology bro, this one will fill any labor shortage bro, our nanas and papas will let them into their homes to help them age in place bro, it’s the future bro, I won’t put guns on them I swear bro, bro
March 6, 2025 at 5:18 PM
Bro, just one more humanoid robot, bro please, this one will be the true general purpose technology bro, this one will fill any labor shortage bro, our nanas and papas will let them into their homes to help them age in place bro, it’s the future bro, I won’t put guns on them I swear bro, bro
This is my feeling as well. It's not quite as mine as peak Twitter was yet but I no longer feel like I'm missing much.
It has now been a full year since I last logged into X, and six months since I last posted on Threads.
Bluesky has become my main source of AI news & discussion.
It has not completely replaced my experience from pre-Elon Twitter (yet!), but it is pretty close.
Bluesky has become my main source of AI news & discussion.
It has not completely replaced my experience from pre-Elon Twitter (yet!), but it is pretty close.
August 1, 2025 at 9:58 AM
This is my feeling as well. It's not quite as mine as peak Twitter was yet but I no longer feel like I'm missing much.
Reposted by Martin Singh-Blom
Lee Mager at LSE posted this about ChatGPT Study Mode and I can't stop laughing
July 31, 2025 at 1:15 AM
Lee Mager at LSE posted this about ChatGPT Study Mode and I can't stop laughing
Reposted by Martin Singh-Blom
If you have infinite data and compute, 1-nearest neighbors is also optimal.
July 31, 2025 at 4:47 AM
If you have infinite data and compute, 1-nearest neighbors is also optimal.
Reposted by Martin Singh-Blom
January 5, 2025 at 10:17 PM
Reposted by Martin Singh-Blom
/
* yes, NWT means capital tax is lower the higher-return is the investment - ie incentivising better investments/faster growth, relative to other ways of taxing capital. It's a bug not a feature. Favour entrepreneurs over rentiers and heirs.
on.ft.com/44FZKhk
/
* yes, NWT means capital tax is lower the higher-return is the investment - ie incentivising better investments/faster growth, relative to other ways of taxing capital. It's a bug not a feature. Favour entrepreneurs over rentiers and heirs.
on.ft.com/44FZKhk
/
Why a wealth tax is capitalism’s handmaiden
[FREE TO READ] Taxing capital holdings boosts rewards for investing well
on.ft.com
July 22, 2025 at 9:58 AM
/
* yes, NWT means capital tax is lower the higher-return is the investment - ie incentivising better investments/faster growth, relative to other ways of taxing capital. It's a bug not a feature. Favour entrepreneurs over rentiers and heirs.
on.ft.com/44FZKhk
/
* yes, NWT means capital tax is lower the higher-return is the investment - ie incentivising better investments/faster growth, relative to other ways of taxing capital. It's a bug not a feature. Favour entrepreneurs over rentiers and heirs.
on.ft.com/44FZKhk
/
Reposted by Martin Singh-Blom
The four horsemen of CEO fear:
July 18, 2025 at 1:11 PM
The four horsemen of CEO fear:
Reposted by Martin Singh-Blom
Reposted by Martin Singh-Blom
"On the Tradeoffs of SSMs and Transformers" by Albert Gu, an excellent read (why is he not on Bsky 😉)
goombalab.github.io/blog/2025/tr...
1/2
goombalab.github.io/blog/2025/tr...
1/2
July 10, 2025 at 8:36 AM
"On the Tradeoffs of SSMs and Transformers" by Albert Gu, an excellent read (why is he not on Bsky 😉)
goombalab.github.io/blog/2025/tr...
1/2
goombalab.github.io/blog/2025/tr...
1/2
Reposted by Martin Singh-Blom
AI slows down experienced developers (against their expectations) in a randomized controlled trial of AI coding effectiveness.
metr.org/Early_2025_A...
metr.org/Early_2025_A...
July 10, 2025 at 7:10 PM
AI slows down experienced developers (against their expectations) in a randomized controlled trial of AI coding effectiveness.
metr.org/Early_2025_A...
metr.org/Early_2025_A...
Reposted by Martin Singh-Blom
Quitting programming as a career right now because of LLMs would be like quitting carpentry as a career thanks to the invention of the table saw.
July 3, 2025 at 2:36 PM
Quitting programming as a career right now because of LLMs would be like quitting carpentry as a career thanks to the invention of the table saw.
Reposted by Martin Singh-Blom
Europe must defend Europe. #SkyShieldNow
July 4, 2025 at 3:46 AM
Europe must defend Europe. #SkyShieldNow
Reposted by Martin Singh-Blom
Netanyahu Calls Iran Strikes Necessary To Prevent War He Just Started
June 13, 2025 at 5:50 PM
Netanyahu Calls Iran Strikes Necessary To Prevent War He Just Started