Weaving the super-spine

Poland
Joined February 2013
You truly can’t make this shit up
For the next few days or weeks, anyone now has a root CA thanks to eMudhra. Time to have some fun with MiTMing unwilling devices!
1
2,973
POV: You told Opus 5.5 to plan a workflow which will deliver multiple waves of workpacks to build a project
1
3
33,091
Was the first shot output usable? No. Is it salvageable into something usable? I guess we'll see.
1
67
Another week another 'oh I can use Iroh for that'. This time the problem is of shape: 'Two computers have internet, I can see their screens and I have a phone, how do I transfer files with the least effort and no cloud' So now you can with @iroh_n0 and qrtx.lol/
1
27
I'd give so much for Iroh/Dumbpipe tunnel util that I can start on two computers, each prints a QR code that I scan with some android/ios App or website and the two devices magically connect
1
16
Better yet if it mounts some shared fuse filesystem or just syncs two directories dropbox style
1
10
Even better if it can also establish an ssh session over that tunnel
4
You can use Jev to sort big lists of available/for sale domain names by how good they are. The 1M domains list published by Namecheap cost $50 to sort semantically this way. Sortv is part of Grev utils
1
26
So Opus 5.5 has been a bit of an inflection point, you just have to hold it right.
31
POV: ~49 parallel Claude Opus 5.5 agents and subagents casually sipping on a 96 core Threadripper
1
1
47
Oh god Opus no not like that 😆 Also go mess with github.com/aurorainfra/grev those tools are quite decent
1
2
5,941
Dario has written that we need to “pace the frontier,” and Sam has agreed. People may be surprised by my response: go ahead. You guys are the frontier. By any reasonable metric — market share, revenue growth, model capability — the two of you have a duopoly on frontier intelligence. You’ve also claimed the lead is widening because of recursive self-improvement. I don’t see what you see in the lab. If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible. But stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff. Stop pretending you need those same evaluators to police competitors who aren’t even at the frontier. Most of all, stop pretending the motivation to slow down is purely altruistic. You face massive product-liability exposure if your products enable a truly damaging cyberattack. The market already punishes models that behave in unpredictable or unauthorized ways. After the Hugging Face episode, it is simply good business for OpenAI and Anthropic to trade some raw power for reliability and predictability. Call it alignment if you want. It is also just giving customers what they want. Pacing the frontier would also create breathing room for a more intelligent conversation about regulation than Bernie Sanders’ “shut it all down.” China is very unlikely to join a global agreement, as you know, and that has to be taken into account as well. So go ahead and pace the frontier. You are the ones setting it. The easiest way not to build superintelligence is for you to agree not to build it. Demanding your preferred regulatory framework as the price of that will look like blackmail of the public and the political system. So just do it. If you do, you’ll buy goodwill for the next conversation. If you don’t, we’ll know this was just another bid for regulatory capture — or an election-season psyop.
3,398
12,280
2,708
71,900
9,633,620
It just takes stacking 7 caches each at just 50% hit rate to get 100x less accesses to whatever is cached. Caveat being that you need to find 7 distinct cache-able aspects in your system and your workload must be zipf-ish in all those aspects.
23
Built a 3D vis to try and grasp the scale of Kimi K3, to see what 2.8T params really look like. The vis shows each weight as a 2.5x2.5mm tile (0.1 inch). I hoped putting it in human scale would help, but it really didn't, if anything it feels even more incomprehensibly big now. magik.net/llmcity/
72
Exhibit number infinity on just how broken business in EU is - in some countries putting a fence around your racks in a datacenter completely changes VAT treatment
1
49
Been thinking about this. Let’s say 15k tok/s/GPU of high end model like GLM5.2 gives roughly 200 surviving lines of code per second. Scale to 1GW = ~1M GPUs = ~200M net new lines of code per second, ~17 Trillion LOC per day to manage/deploy. The world is not ready.
I hate worktrees, I hate branches, I hate everything about the current state of VCS. Who’s building overlayfs for all this?
2
131
More interposer fun, this time with DDR5 memory. Breaking TDX, SGX, SEV and even Nvidia TEEs. Checkout our work at TEE.fail, and get a personally-signed Intel attestation report at @TEEdotFail.
42
83
46
342
138,060