Sign in

Martin Singh-Blom

@singhblom.com
1.4K followers 1.3K following 220 posts

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.

PostsRepliesMedia
Reposted by Martin Singh-Blom
Tom Silver @tomssilver.bsky.social · 13/09/2026
This week's #PaperILike is "Latent Programming Horizons in Coding Agents" (Silva et al., 2026). Now seems like a good time to better understand what coding agents are doing. Here's a nice example of the kind of analysis one can do. PDF: arxiv.org/abs/2607.05188
arxiv.org
Latent Programming Horizons in Coding Agents
A coding agent solving a software-engineering task spends dozens of steps reasoning, editing code, and running tests, yet little is known about what the underlying language model internally represents...
084
Reposted by Martin Singh-Blom
New-Cleckit Dominie @ncdominie.bsky.social · 09/09/2026
I see that OpenAI's messaging of "if you give us all the money on Earth we will annoy some mathematicians" is being somewhat overshadowed by Anthropic's messaging of "if you give us all the money on Earth we will kill everyone". I'm slightly worried about what a compromise position might look like.
45817
Reposted by Martin Singh-Blom
mr. TIM @timkellogg.me · 08/09/2026
it’s official: OpenAI claims they solved Navier-Stokes 10,000 concurrent agents coordinated by an internal model (that’s currently still in training) worked for 88 hours openai.com/index/navier...
We're sharing a solution to the Navier-Stokes existence and smoothness problem,
one of the Millennium Prize Problems. This proof, produced by an internal OpenAl system, shows that the dynamics of the Navier-Stokes equations for fluid motion can develop a singularity in finite time. We're sharing both a writeup of the proof and a formalization in Lean.
75410
Martin Singh-Blom @singhblom.com · 08/09/2026
The OpenAI Navier-Stokes drama is good drama. Did OpenAI train on conversations where mathematicians were working on Clay prize proofs? What should that mean for credit assignment ?https://news.ycombinator.com/item?id=49605915
news.ycombinator.com
Navier-Stokes – Tristan Buckmaster [pdf] | Hacker News
001
Martin Singh-Blom @singhblom.com · 03/09/2026
This was really nice. Pair programming during the planning phase with a colleague and with the LLM seems like something that should be done much more. Align on architecture before the clanker starts clanking.
010
Reposted by Martin Singh-Blom
Ed @ed3d.net · 30/08/2026
GLM-5.3-Flash rips the doors off and is under a full MIT license we have, and I mean this literally, a Smart Enough model for almost any task a normal person and many AI perverts can envision, and it is free to use at the same time, inference hardware will only get more accessible to you
740745
Martin Singh-Blom @singhblom.com · 12/08/2026
Trying to maximize the science time with the kids this evening. The eclipse had them super excited.
Eclipse seen through a telescope, projected onto a piece of paper.
020
Reposted by Martin Singh-Blom
LOWρUFO 🇺🇦🇵🇸🌍 @lowrhoufo.bsky.social · 07/08/2026
Absolutely dying lmao
1341383
Reposted by Martin Singh-Blom
Dave Keating @davekeating.substack.com · 31/07/2026
There's growing evidence the surge into Ceuta (🇪🇸PM Sanchez has called it an "attack") has, like in 2021, been orchestrated by the Moroccan government in retaliation for Sanchez visiting Algeria Yet it's Spain being criticised by European politicians, not Morocco. So much for solidarity.
bbc.com
About 60,000 migrants arrive in Ceuta in 24 hours, Spanish territory's leader says
The estimated number of arrivals is more than 70% of the population of the city, which is about 83,600.
24496284
Reposted by Martin Singh-Blom
Dashiell @dashiells.bsky.social · 28/07/2026
AI safety folks should be way more worried about concentration of power than the proliferation of powerful models
37710
Reposted by Martin Singh-Blom
Thorne 🌸 @ens0.me · 24/07/2026
It's wayyyyy better for society if we have high-quality open weight models where you choose to run inference. And if China's the one to do it, so be it.
2703
Reposted by Martin Singh-Blom
Jay Rosen @jayrosen.bsky.social · 20/07/2026
Here's something I post from time to time. My answer to a reader who asked me: what could journalists do NOW to break with some of their more corrosive habits.
361836591
Reposted by Martin Singh-Blom
Gingerspiced @gingerspiced.bsky.social · 18/07/2026
Pete Hegseth: "We need our warriors to embrace the Spartan mindset" U.S. Military: *suffers crushing defeat to the Persians*
317254956964
Reposted by Martin Singh-Blom
Atlas Shvitzed @therafischwartz.bsky.social · 08/07/2026
Unfortunately, a stunned Paris is how they all got into this whole mess to begin with
360159473256
Reposted by Martin Singh-Blom
hardmaru @hardmaru.bsky.social · 22/06/2026
Human intelligence is fundamentally a collective intelligence. We solve complex problems by participating in a vast cultural network that builds upon ideas across generations. I believe the strongest AI systems will become a collective intelligence, too.
77111
Reposted by Martin Singh-Blom
Sky Marchini @sky.skymarchini.net · 20/06/2026
uhm aiclambake.com/clamtakes/li...
aiclambake.com
AI Engineer Claims to Have Cracked Linear A
AI Engineer Claims to Have Cracked Linear A
2221623
Reposted by Martin Singh-Blom
peterdoohan.bsky.social @peterdoohan.bsky.social · 11/06/2026
How do brains plan actions towards goals? To get at this question we studied mice navigating complex mazes as goals changed on every trial 🧵 Work with @thomasakam.bsky.social @behrenstimb.bsky.social @kristorpjensen.bsky.social now on BioRxiv: www.biorxiv.org/content/10.6...
411846
Martin Singh-Blom @singhblom.com · 04/05/2026
15 best movies quickly: In the mood for love The seven samurai Amelie No country for old men Inglorious Basterds Mother of Mine Raise the red lantern Bridge over river Kwai The Martian Good Will Hunting Die Hard Crouching Tiger, Hidden Dragon Life of Brian La Vita e Bella Death of Stalin
010
Reposted by Martin Singh-Blom
Ryan Moulton @moultano.bsky.social · 01/05/2026
Generational tweet that might even spawn a school of architecture.
Cairo Smith:
There's a common misconception that Brutalist buildings were unpainted, but thanks to microscopic analysis of the exteriors we can now recreate what they looked like in their prime.
17494481745
Reposted by Martin Singh-Blom
J. Craig Venter Institute @jcvi.org · 30/04/2026
It is with great sadness that we announce J. Craig Venter has died. 🧪🧬 www.jcvi.org/media-center...
jcvi.org
J. Craig Venter, genomics pioneer and founder of JCVI and Diploid Genomics, Inc., dies at 79
La Jolla, California—April 29, 2026—The J. Craig Venter Institute (JCVI) announced that J. Craig Venter, Ph.D., the Institute’s founder, board...
3187123
Reposted by Martin Singh-Blom
Tom Silver @tomssilver.bsky.social · 05/04/2026
This week's #PaperILike is "AssistanceZero: Scalably Solving Assistance Games" (Laidlaw et al., ICML 2025). AlphaZero-like combo of learning & planning for assistance games, where robot & human share reward fn that robot doesn't know. + Minecraft! PDF: arxiv.org/abs/2504.07091
arxiv.org
AssistanceZero: Scalably Solving Assistance Games
Assistance games are a promising alternative to reinforcement learning from human feedback (RLHF) for training AI assistants. Assistance games resolve key drawbacks of RLHF, such as incentives for dec...
0194
Reposted by Martin Singh-Blom
Steven Nguyen Scaife @midfalutin.bsky.social · 05/04/2026
have been researching bestiaries for a project and found this one of a griffin about to tear off some poor soul's entire head very relatable
a medieval bestiary page depicting a griffin on a person's back with their entire head in its mouth
3368
Reposted by Martin Singh-Blom
Allison Gill @muellershewrote.com · 05/04/2026
Happy Easter.
32280381235
Reposted by Martin Singh-Blom
Christian Breyer @christianonre.bsky.social · 04/04/2026
Global solar additions reached record 511 GW [613 GWp with 1.2 Wp/W,ac ratio] in 2025, says IRENA www.pv-magazine.com/2026/04/02/g... - and much more ahead: doi.org/10.1016/j.en... ; doi.org/10.1016/j.rs...
pv-magazine.com
Global solar additions reached record 511 GW in 2025, says IRENA
Latest figures from the International Renewable Energy Agency find solar contributed the majority of a record 692 GW of renewables capacity added worldwide last year.
05417
Reposted by Martin Singh-Blom
Ethan Mollick @emollick.bsky.social · 04/04/2026
A new field experiment on 515 startups, half shown case studies of how startups are successfully using AI. Those firms used AI 44% more, had 1.9x higher revenue, needed 39% less capital: 1) AI accelerates businesses 2) The challenge is understanding how to use it papers.ssrn.com/sol3/Deliver...
2737
Reposted by Martin Singh-Blom
Jonathan Stray @jonathanstray.com · 27/03/2026
Could social media make us less polarized instead of more? We tested 5 algorithms on 3 platforms with 10,000 people for 6 months during the 2024 election, and found that the answer is yes. 🧵
29230
Reposted by Martin Singh-Blom
Evan Peck @peck.phd · 18/03/2026
I made a tiny tool for quickly sharing small datasets (< ~1000 rows) without uploading any data to a server. 🔗 ziptbl.com It compresses the data into the link itself, so there’s no account, hosting, or storage layer involved. Here's Florence Nightingale's famous 📊 data: ziptbl.com#d=eNpdlE-LGz...
1430996
Martin Singh-Blom @singhblom.com · 14/03/2026
This is the funniest optimization method I've seen in a long time. Really cool work.
081
Reposted by Martin Singh-Blom
Jeremy Berg @jeremymberg.bsky.social · 01/03/2026
NSF Update (Awards through 2/27/26) Directorates to follow 1/10
A line graph showing NSF grant awards made through 2/27/26 for fiscal year 2026 compared with grant awards for fiscal years 2021-2025.
30675445
Reposted by Martin Singh-Blom
Claus Wilke @clauswilke.com · 09/02/2026
Just a regular reminder that there is absolutely no scientific evidence that caffeine (in moderation) is in any way problematic. Caffeine is this thing that you think should be bad for you but it's just not. Most studies suggest it has meaningful benefits. See reply for a recent example.
2112
Reposted by Martin Singh-Blom
Nikhil Garg @nkgarg.bsky.social · 14/01/2026
If you enjoy Paper Skygest, you might enjoy our new academic preprint describing it and our vision! arxiv.org/abs/2601.04253 bsky.app/profile/pape...
1315
Reposted by Martin Singh-Blom
Una Hajdari @unawrites.bsky.social · 05/01/2026
Shamelessly stolen from the other website, because of course that specific part of the crisis would have been averted if Europeans were involved 🤣🤣🤣
0142
Reposted by Martin Singh-Blom
Dmitry Grozoubinski @explaintrade.com · 04/01/2026
Apart from everything else, invading a country for oil in 2026 is somewhat like doing so for horses in 1931.
1746099
Martin Singh-Blom @singhblom.com · 03/01/2026
ARC Raiders won Most Innovative Gameplay in the 2025 Steam Awards! It has been really fun to work on and my coworkers are so amazingly creative and clever. Really nice to see that their cleverness comes through in how players perceive the game!
020
Reposted by Martin Singh-Blom
Mark Chadbourn @chadbourn.bsky.social · 03/01/2026
After this, no one should doubt Trump’s intention to invade Greenland. The EU needs to step up fast.
28981315
Reposted by Martin Singh-Blom
Gavin Chait; @gavin.whyqd.com · 03/01/2026
And, just like that, an era of international collaboration and rule of law under the United Nations, ends. What a pathetic and weak-willed response. Accepting American violence means that EU interests won't be defended, and the world is infinitely more dangerous.
1188
Reposted by Martin Singh-Blom
Mike Black @mikeblack114.bsky.social · 03/01/2026
Simple as Everyone in DoD who participated in planning and executing this needs to face charges Yes I know the scope/scale of what I'm saying; either laws mean things or they don't
161436381
Reposted by Martin Singh-Blom
Olena Halushka @halushka.bsky.social · 02/01/2026
Horrifying photos from Kharkiv, where russian missiles have just wiped out a residential building. At least 13 people were injured. I am afraid the figures will be worse. Please don't look away. This is not normal. russian terrorism is not normal.
817287
Reposted by Martin Singh-Blom
Olúfẹ́mi O. Táíwò @olufemiotaiwo.bsky.social · 02/01/2026
funding lotteries have been waiting for this moment
422961
Reposted by Martin Singh-Blom
Michael Friendly @datavisfriendly.bsky.social · 02/01/2026
ggCheysson is based on colors, patterns of these _Albums_ maps collected by @infowetrust.com Cheysson fonts by @kennethfield.bsky.social
0298
Reposted by Martin Singh-Blom
Michael Spicer @michaelspicer.bsky.social · 30/12/2025
how all youtube ads sound to me
621571443
Reposted by Martin Singh-Blom
Seva @seva.bsky.social · 30/12/2025
I always tell students if you ask a good question the article will write itself
322531515
Reposted by Martin Singh-Blom
Alucard (悪魔城下町) @akumajoukamachi.bsky.social · 29/12/2025
Important and powerfully written piece. Some excerpts: "...once a state grants itself the power to unmake a citizen, the line can always be moved." "Sweden, now with a population of 10 million, has the world’s highest proportion of dollar billionaires per capita."
1439
Martin Singh-Blom @singhblom.com · 07/12/2025
Check out this really fun episode where @dannyodwyer.bsky.social interviews a bunch of Embarkers, including me, about how hard it is to train robots when they don't do what you want them to do!
020
Reposted by Martin Singh-Blom
Chris Paxton @cpaxton.bsky.social · 26/11/2025
Why not everything in manufacturing is automated (yet) Every few weeks I see someone wondering why so little of manufacturing, particularly in the United States, is automated. There are tons of reasons why, but it's still very difficult to set up robots to succeed: open.substack.com/pub/itcanthi...
2417
Reposted by Martin Singh-Blom
Five Minute Macro @fiveminutemacro.bsky.social · 11/10/2025
Incredible work by the Dallas Fed.
3407
Reposted by Martin Singh-Blom
Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 08/10/2025
Soooooo when is bluesky deploying bridging systems arxiv.org/abs/2301.09976
arxiv.org
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...
1193