1.
Air-gapped exfiltration through low-frequency magnetic fields
Low-frequency magnetic fields can pass through Faraday cages, giving attackers a plausible side channel even against shielded air-gapped machines
Backlist — 10 May 2026 UTC
Balanced concrete AI systems work against security, semiconductors, infrastructure, science policy, markets, health, architecture, graphics, and language to avoid an agent-only slate.
2.
Building a GPT-2 inference engine from scratch in CUDA
Implementing transformer inference directly in CUDA exposes the real performance work behind reductions, memory behavior, numerical stability, and kernel design
3.
Universities sued over ImageNet copyright claims
A lawsuit over ImageNet could test how fair use applies to foundational academic datasets, not just commercial model training
4.
A Georgia data center used 29M gallons of water before anyone billed it
The Fayetteville case shows how data-center demand can stress local utilities before billing, monitoring, and governance systems catch up
5.
The bottleneck was not microelectronics, it was U.S. composite-hull manufacturing
A defense-hardware startup had to set up production in Turkey because U.S. shipyards could not scale carbon-composite hulls at 10–20 units per year
6.
Private collectors now own more T. rex specimens than public institutions (t.co)
Wealthy buyers are turning important dinosaur fossils into private assets, reducing the material available for reproducible paleontology research
7.
SMEE’s 28nm ArFi lithography tool reportedly enters mass production
A Chinese lithography tool capable of 28nm single exposure and more advanced multi-patterning would materially change the constraints on domestic chipmaking
8.
PyTorch now has public developer logs (x.com)
Meta is moving internal build notes into public PyTorch devlogs, making implementation decisions and maintenance work easier for outsiders to follow
9.
Redis persistence via append-only files
Redis is widely treated as an ephemeral cache, but its append-only file mode provides a durable log of writes for crash recovery and replay
10.
vx-underground adds roughly 200k malware samples (t.co)
A large public malware archive gives defenders, researchers, and tooling authors more real samples to test against instead of relying on toy datasets
11.
MindOn’s Unitree G1 demo shows a full autonomy stack on humanoid hardware
The claimed pipeline combines whole-body control, contact-rich planning, embodied action sequencing, and onboard perception on commodity humanoid hardware
12.
Simulating drone-detector antenna placement
Antenna placement changes detection range across frequencies enough that drone defense depends on geometry, not just sensor quality
13.
An interactive map for accessing U.S. state court records (x.com)
A crowdsourced guide to state court access became a map that legal journalists can use to find records without rediscovering each jurisdiction’s rules
14.
Regulating access to platform data (x.com)
Platform-data regulation has to balance competition against the incentive for platforms to create and maintain valuable data in the first place
15.
India’s GCCs reach $100B in revenue
Global capability centers in India now employ 2.3M people and generate more revenue than the country’s Big Four IT services firms combined
16.
DFlash enters a production inference stack (t.co)
A draft-model system moving into production signals that speculative decoding and related inference optimizations are becoming operational infrastructure
17.
Why fine-tune in 2026?
Fine-tuning still buys product teams tighter UX control and better cost-speed-quality tradeoffs than prompting frontier models alone
18.
Notes on the Polymarket V2 API changes
Polymarket’s V2 API moves toward shared attribution slots and simpler operator-managed payloads, a pattern other prediction markets may copy
19.
Hyperscaler cash flow during the AI capex buildout
Forecasts show AI capital spending pushing most hyperscalers near zero or negative free cash flow while Microsoft remains the relative outlier
20.
Memory chip shortage may last until 2027
Samsung, SK Hynix, and Micron are prioritizing AI demand, and new memory capacity takes long enough that shortages may persist for years
21.
Illusions of Understanding in the Sciences (t.co)
Examples from bicycle stability to aerodynamic lift show how scientific communities can overestimate how well they understand familiar phenomena
22.
The Fall of the Theorem Economy (x.com)
David Bessis argues that AI changes the economics of mathematical proof and may expand who can participate in serious mathematics
23.
What American IVF clinics often do not tell patients
Egg-freezing and IVF protocols vary widely, and the default U.S. approach is often more aggressive than the alternatives used elsewhere
24.
A temporary shrine roof grows into a forest before dismantling
The Dazaifu Tenmangu temporary shrine roof was designed to mature into the landscape over three years and then disappear
25.
The Mahabharata rendered as 13,514 WebGPU pages in a Fibonacci sphere
Rendering an entire Sanskrit epic as GPU text on a sphere turns typography, shaders, and large-document visualization into one experiment
26.
Testing superconducting magnetic shielding in space
A radiation sensor aboard the Almighty Magnet mission will test whether ultra-strong superconducting magnets can reduce radiation exposure beyond Earth
27.
Sentinel-2 shows tankers burning off Jask
Satellite imagery confirmed multiple oil tankers burning near the Iranian port of Jask, grounding regional conflict claims in observable evidence
28.
HTML Drive: versioned storage for agent-generated HTML (t.co)
Agents that create HTML can commit diffs to a minimal drive with automatic versioning, borrowing the useful parts of Google Drive and Git
29.
Jack Clark’s case for AI automating AI research by 2028
The argument assigns a 60% chance that frontier models can autonomously train successor models by the end of 2028, based on progress in software engineering and research workflows
30.
The German phrase for when objects have it in for us
Die Tücke des Objekts captures a common experience English lacks a compact idiom for: the perverse resistance of things
31.
1 GW AI data center economics are mostly a sensitivity table based on the assumptions you use.
1 GW AI data center economics are mostly a sensitivity table based on the assumptions you use. Three revenue cases: Low: ~$7B (500k GPUs × 80% utilization × $2/GPU-hour × 8,760 hours) Mid: ~$17B (750k GPUs × 85% utilization × $3/GPU-hour
32.
it's essentially overwhelmingly one guy who was responsible for reverse engineering and reimplementing the ninten…
it's essentially overwhelmingly one guy who was responsible for reverse engineering and reimplementing the nintendo switch's entire os kernel over the course of several years
33.
This library (kill-port) has 1,4M weekly downloads on npm and it usually takes ~10s to kill a process in MacOS.
This library (kill-port) has 1,4M weekly downloads on npm and it usually takes ~10s to kill a process in MacOS. So... I rewrote it in Rust and now it takes 3ms That's just 3,000x faster than the original. Been using it for a couple week
34.
My "Hello World" for new model/harness is building a Lisp interpreter in Rust and one in Python. Guess which one …
My "Hello World" for new model/harness is building a Lisp interpreter in Rust and one in Python. Guess which one nailed both?
35.
funny you say that, been building a 64-bit RISC-V emulator in Rust for the past few days.
funny you say that, been building a 64-bit RISC-V emulator in Rust for the past few days. time-travel debugger included, you can seek to any cpu cycle :)
36.
Docs from (x.com)
Docs from @NousResearch on how to set up Pareto Code in Hermes: https:// hermes-agent.nousresearch.com/docs/user-guid e/configuration#openrouter-routing--pareto-code-for-auxiliary-tasks …
37.
DS4 running on DGX Spark (GB10 / CUDA), private branch for now. 12 tokens/sec, the memory bandwidth is limited in…
DS4 running on DGX Spark (GB10 / CUDA), private branch for now. 12 tokens/sec, the memory bandwidth is limited in this system, at 270GB/sec. But prefill is ways more alighed to M3 Max at ~200 t/s. I'll release when more mature, but it is al
38.
Codex iterated a pure NumPy + cv2 closed-loop heuristic policy for VizDoom D3 Battle. No neural network training,…
Codex iterated a pure NumPy + cv2 closed-loop heuristic policy for VizDoom D3 Battle. No neural network training, no map, no object coordinates, no seed-specific routes. Just screen pixels plus public game variables, roughly the same signal
39.
releasing hf-sandbox
releasing hf-sandbox
40.
If I extract the analysis channel, I can see how GPT5.5 sometimes reframes my question to deny having influences
If I extract the analysis channel, I can see how GPT5.5 sometimes reframes my question to deny having influences (And, yes, I vibe-coded a tool that extracts GPT-5.5's CoT via prompt injection)
41.
Modded-NanoGPT optimization result #12: Transferring good hparams from recent NorMuon records -- in particular, t… (x.com)
Modded-NanoGPT optimization result #12: Transferring good hparams from recent NorMuon records -- in particular, taking final val 25 steps early following @wen_kaiyue 's NorMuonH, and lr=0.035 following Liming Liu's NorMuon -- improved the
42.
tilde research just found a massive flaw in the muon optimizer powering deepseek v4 and kimi k2.5
tilde research just found a massive flaw in the muon optimizer powering deepseek v4 and kimi k2.5 turns out muon permanently kills over 25% of your mlp neurons in early training so they built aurora to fix it and the benchmarks are actuall
43.
One thing I need to have a PC setup and running is having codex port SwiftUI to windows
One thing I need to have a PC setup and running is having codex port SwiftUI to windows I have a nice starting point on it but I decided I want ut to go low level Write its own render engine direct to render APIs on windows
44.
preview of the ethereum lending dashboard
preview of the ethereum lending dashboard -aave, spark, morpho, fluid, euler coverage -vaults, rates, allocations, flows, risks -history for all the individual positions -unlimited api access for real-time monitoring everything is ready,
45.
1. automatically add paths referenced in prompt to the permissions
1. automatically add paths referenced in prompt to the permissions 2. add "async permissions". instead of waiting in the tool forever return a message right away: "permission for this tool call is suspended. the user is not at the compute
46.
"Welding Operations in Hazardous Locations Using Humanoid Robots (VR Remote Operation)" (t.co)
"Welding Operations in Hazardous Locations Using Humanoid Robots (VR Remote Operation)" https:// tv.cctv.com/2026/04/12/VID EIiaf7vQ1VmqyNyXdDKIL260412.shtml … #humanoidrobot #teleoperation #industrial #welding #infrastructure #maintenanc
47.
Well Code cooked, Doom in Swift is almost a 100% accurate rendering now. The engine builds a paletted 320x200 fra…
Well Code cooked, Doom in Swift is almost a 100% accurate rendering now. The engine builds a paletted 320x200 framebuffer in Swift. The macOS shell only presents that finished framebuffer in a native window
48.
cloudflare’s actually cooking
cloudflare’s actually cooking i used to use it purely for domains but i’ve pretty much started using it + planetscale for everything, especially because AI agents make it very easy to get stuff set up
49.
codex seems to have full source access and still can't get the BSP renderer right after 40 hours :/
codex seems to have full source access and still can't get the BSP renderer right after 40 hours :/ nothing in the original sources is tricky. a straight port is pretty trivial and mostly mechanic. and yet.
50.
one of the tricky things about the rust port is layering. it’s currently many dozens of crates, which speeds up c…
one of the tricky things about the rust port is layering. it’s currently many dozens of crates, which speeds up compile times but blocks cyclic dependencies. a lot of bun’s zig codebase uses tagged pointers for interfaces, for things like
51.
Can’t believe it but I’ve turned this concept into a functional iOS app with the magic of Codex
Can’t believe it but I’ve turned this concept into a functional iOS app with the magic of Codex 8 little agents powered by Apple Foundation models with customizable system prompts in an iMessage-style UI
52.
Deepseek has all my respect as they own almost every corner of their tech stack, from recipes, training framework…
Deepseek has all my respect as they own almost every corner of their tech stack, from recipes, training framework to kernels. One common thing for telling a frontier organization is whether it treats software sovereignty for getting quick
53.
When it comes to fighting compatibility issues on GB200 (90% of what I do for the past 2 months), I might just bu…
When it comes to fighting compatibility issues on GB200 (90% of what I do for the past 2 months), I might just buy the farm somewhere remote and start grazing sheep
54.
I'm working on a new android launcher for my phone and it was a little laggy so I just told Codex "please make it…
I'm working on a new android launcher for my phone and it was a little laggy so I just told Codex "please make it as snappy and fast as possible, 0ms latency when I swipe up to go home" and 20 minutes later there is 0ms of latency when I sw
55.
A few people were asking how to bring their Codex pets onto hardware devices, so I made a walkthrough of how to f…
A few people were asking how to bring their Codex pets onto hardware devices, so I made a walkthrough of how to flash pets using the Badge As promised, the github repo for the integration with the Codex App is in the comments Sharing the
56.
Yesterday, (x.com)
Yesterday, @Storyaliz had our first outage! @neondatabase had an outage and our DB was down. This is not a milestone I was looking forward to, but it is a milestone all the same. And, good AND bad, users were online and impacted.
57.
ICYMI, looots of new tutorials landed in OpenEnv docs. (t.co)
ICYMI, looots of new tutorials landed in OpenEnv docs. go get started with RL envs! https:// meta-pytorch.org/OpenEnv/tutori als/index.html …
58.
Nonsense helps LLMs reason better
Nonsense helps LLMs reason better LoPE prepends Lorem Ipsum to prompts when GRPO hits the zero-advantage problem, unlocking orthogonal reasoning paths and boosting math scores across 1.7B-7B models.
59.
Great technical long post! Very bullish power semis and testers. Ohm’s law FTW!
Great technical long post! Very bullish power semis and testers. Ohm’s law FTW! “Power semiconductor content per rack grows substantially across this transition. SiC and GaN suppliers, high-voltage busbar and connector vendors, and rack-le
60.
If you look closely you can see how Waymo is tracking that car from way before you can see it in the camera view …
If you look closely you can see how Waymo is tracking that car from way before you can see it in the camera view (roof mounted LiDAR stays winning), and how quickly the trajectory starts bending the moment it becomes obvious that the human
61.
We are measuring directionally similar, but even more striking difference: 5.5 is a better base model, but the dr…
We are measuring directionally similar, but even more striking difference: 5.5 is a better base model, but the drastically reduced thinking budget (at the same xhigh) makes it worse for high-complexity tasks, like bug finding. We need to be
62.
Anthropic’s recent interp work is awesome. A few months ago, I felt strongly that AI companies needed to make fas…
Anthropic’s recent interp work is awesome. A few months ago, I felt strongly that AI companies needed to make faster progress understanding *why* models engage in behaviors researchers tried to prevent. And they’re making progress faster th
63.
blown away by how LR insensitive PSGD is
blown away by how LR insensitive PSGD is
64.
Built out a yolo /remote-control in the Codex cli using /goal. (x.com)
Built out a yolo /remote-control in the Codex cli using /goal. - /remote-control starts a tiny server on laptop - generates fresh token and qr code - phone connects through webapp - full sync between phone and laptop codex - touch grass A
65.
This is a really cool paper on Latent Action Models and provides cool ideas of how can we evaluate action represe…
This is a really cool paper on Latent Action Models and provides cool ideas of how can we evaluate action representations in latent space
66.
USDAI is a financing vehicle for the AI capex boom: a tradable, GPU-backed debt product onchain. (x.com)
USDAI is a financing vehicle for the AI capex boom: a tradable, GPU-backed debt product onchain. - USDai: the stablecoin, used for payments like loan settlement and interest payments - sUSDai: the yield product, used to fund the AI buildou
67.
last month i wrote a blog on memory internals of hermes-agent by (x.com)
last month i wrote a blog on memory internals of hermes-agent by @NousResearch thought i should share it here https:// samyak1729.github.io/hermes-blog/
68.
some thoughts on the shape of foundation labs
some thoughts on the shape of foundation labs 1) epoch ai estimated anthropic @ $9m in revenue per employee and openai @ 5.6m in revenue per employee 2) these rates would be the highest among public technology companies; but, i'm not sure
69.
> got codex pro 20x
> got codex pro 20x > burnt 97% weekly limits > generated 107M dataset > fine-tuned a 4B model > beaten sonnet 4.6 by 23% > no regrets!
70.
This is awesome! This behavior is exactly what we benchmark in (t.co)
This is awesome! This behavior is exactly what we benchmark in http:// CodeClash.ai where LMs play against each other in 7 different arenas by writing code. I think there's *so* much more to do in this research direction, and the impacts w
71.
We are hiring research fellows to help us improve FrontierSWE!
We are hiring research fellows to help us improve FrontierSWE! If you want to help build the hardest real-world coding benchmark, reach out! Fellows can work with us for a few weeks up to months and will be supported with compute and a gen
72.
chicken and egg in event markets: better oracles let you issue the long tail of contracts, but you need contracts…
chicken and egg in event markets: better oracles let you issue the long tail of contracts, but you need contracts to exist for oracles to converge on an entire world of outcomes isn't represented in any issued market today - that gap is wh
73.
Tried /goal for the first time. Just threw this challenge at it (t.co)
Tried /goal for the first time. Just threw this challenge at it https:// optimizationarena.com/prop-amm (s/o @gnarayan ) It climbed avg edge from +470 to +510 in 25 hours Cool to throw these auto-research type problems at /goal and se
74.
how do you guys see the solution to permission approval?
how do you guys see the solution to permission approval? 1. have some sort of external notification to approve/deny request 2. have an agent determine if it should be allowed what is your ideal solution to this
75.
Appreciate Ivan tweet. To put this into context, to build DS4 I used: my MacBook M3 Max (mine, 8k euros), 1 M3 Ul…
Appreciate Ivan tweet. To put this into context, to build DS4 I used: my MacBook M3 Max (mine, 8k euros), 1 M3 Ultra with 512 GB (got access, 10k euros), one DGX Spark (got access, 4k euros?). Are we far from the times all you needed to do
76.
@saturdayrobotic (x.com)
@saturdayrobotic Robotics & World Model Reading Club 07 Recap, keynote Ahmet Şemi ASARKAYA ( @agilityrobotics ), hosts @junfanzhu98 , @aurorafeng_01 . DreamerV4: 1.6B Diffusion-Transformer World Model achieves offline Minecraft Diamond
77.
Quick update: Cardputer thing ended up adding Managed Agents ( really cool stuff from the team (x.com)
Quick update: Cardputer thing ended up adding Managed Agents ( really cool stuff from the team @bcherny @ClaudeDevs ) - now you can fire off a Claude agent - it pages you back when the agent finishes - mirror everything in a HT
78.
MirrorCode: You can port programs with proper setup and $$$
MirrorCode: You can port programs with proper setup and $$$ Jarred: You can port programs with proper setup and $$$ Some benches: Models are unable to port programs, it’s a very hard task and they score basically 0
79.
For the first Codex Community Event in London, what type of event would people prefer?
For the first Codex Community Event in London, what type of event would people prefer? Feel free to add other suggestions in thread - I want to make the best event for Agentic Engineers possible.
80.
For 100% agent-written frontends, I keep coming back to this:
For 100% agent-written frontends, I keep coming back to this: Maybe we don't start with a frontend framework Maybe we start with an index.html, browser primitives, Web Components for reusable UI and a strict convention for how agents rout
81.
guess how fun it is having all of the openclaw user base beat up pi's llm provider abstraction.
guess how fun it is having all of the openclaw user base beat up pi's llm provider abstraction. guess i'm "one of the very few teams that have dealt with the quirks between providers at scale" now ...
82.
Terminating and backward process do language server in VsCode is hard. It doesn’t even terminate cpp lang servers
Terminating and backward process do language server in VsCode is hard. It doesn’t even terminate cpp lang servers
83.
how to raise from me as an ai founder:
how to raise from me as an ai founder: dont tell me your model is better. that usually means your business dies the second base models move. tell me what workflow you own that customers cannot rip out even when the intelligence gets cheap
84.
lovely article going deeper into the RL-SFT-OPD spectrum with some very nice intuitions + experiments :)
lovely article going deeper into the RL-SFT-OPD spectrum with some very nice intuitions + experiments :)
85.
Bullish on Demis and the DeepMind team.
Bullish on Demis and the DeepMind team. Beautiful result from GoogleDeepMind: (this one has been quite useful to me) Its an ablation on how the chat template can be an extraction plane on open-weight models. Prompting with just the templat
86.
if you exclude make a custom, reasoning-only ECI (basically all benches but SimpleQA and HLE), Kimi K2.6 is on pa…
if you exclude make a custom, reasoning-only ECI (basically all benches but SimpleQA and HLE), Kimi K2.6 is on par with Gemini 3.0 (frontier 5 months before K2.6) and Muse Spark (released same month) OpEn SoUrCe Is FaLLiNg bEhInD
87.
Brothers! I promised you I'd open-source it tonight, and here it is (t.co)
Brothers! I promised you I'd open-source it tonight, and here it is image to 3D model currently only integrates with the online one: http:// tripo3d.ai You can also modify it for other providers, or local models Don't forget to like and
88.
We now have video proof generation for issues on OpenClaw as part of working on QA automation. (x.com)
We now have video proof generation for issues on OpenClaw as part of working on QA automation. Codex [or a GH workflow] generates before/afters (crabbox does the screen recording). Kudos to @obviyus for automating real Telegram login!
89.
Someone please rewrite rust in rust with the /goal to make compile times faster.
Someone please rewrite rust in rust with the /goal to make compile times faster.
90.
Wrapper startups are toast because frontier lab enterprise sales solves the problem of knowledge diffusion way fa…
Wrapper startups are toast because frontier lab enterprise sales solves the problem of knowledge diffusion way faster than anyone thought