debug

1.3K posts

debug

debug

@0xdbug

Security engineer, developer and hobbyist reverse engineer in that order. Opinions/rants are all mine and not my employer's.

Washington, USA Katılım Mayıs 2009
844 Takip Edilen758 Takipçiler
debug
debug@0xdbug·
Happy pi day you all
English
0
0
2
64
debug
debug@0xdbug·
Comparison between nuclear tech and chip industry is cool. But a big unspoken constraint on nuclear tech is restrictions that govts put aswell. That restriction doesn’t exist for chip and computing. It exists but it’s mostly economic in nature.
Big Think@bigthink

Semiconductors 101 | Chris Miller @crmiller1

English
0
0
0
87
debug retweetledi
Dmitry Vyukov
Dmitry Vyukov@dvyukov·
syzkaller/syzbot now has AI agentic framework for kernel bug fix generation, bug assessment, security triage, POC generation, etc: groups.google.com/g/syzkaller/c/… Includes set of tools to build kernels, navigate/edit source, test reproducers, etc. Contributions/research are welcome.
English
1
39
127
10.9K
debug retweetledi
Chris Lattner
Chris Lattner@clattner_llvm·
One not very hot take - The Claude C Compiler has the best internal architecture docs of any compiler I've ever seen. Far, far, better than any compiler I've ever written, lol :-)
English
14
53
1.1K
79.7K
debug retweetledi
Halvar Flake
Halvar Flake@halvarflake·
It's kinda fascinating that you could never get away shipping CPUs nearly as unreliable, but for GPUs it seems it's OK. Is this also the case for AMD GPUs and TPUs? Or is this just Nvidia building really close to the edge?
SemiAnalysis@SemiAnalysis_

@modal They conclude with: "It’s underappreciated how unreliable GPUs are. NVIDIA’s hardware is a marvel, the FLOPs are absurd. But the reliability is a drag." We agree! (6/9)

English
14
6
44
9K
debug retweetledi
Ryan Dahl
Ryan Dahl@rough__sea·
"The RISC-V port of V8 has come a long way in the last few years. While there is still work to be done, the port is now mostly at feature parity with the officially supported architectures. V8 on RISC-V now runs the full JetStream benchmark suite, which consists of ~33 MB of Wasm bytecode and ~2M lines of JavaScript code – and it is ready for your workloads too." riseproject.dev/2025/12/09/a-g…
English
2
10
117
40.3K
debug retweetledi
Björn Töpel
Björn Töpel@bjorntopel·
The @risc_v devroom for @fosdem got accepted! \o/ Stay tuned for the CfP!!!
English
1
1
7
373
debug retweetledi
Colin Percival
Colin Percival@cperciva·
If every company using FreeBSD in EC2 spent 0.1% of their EC2 spend on sponsoring FreeBSD, it would barely be noticeable in their budgets but it would be a tremendous boon to FreeBSD development. Seriously, FreeBSD is free, but is 0.1% an unreasonable amount to ask for?
English
17
39
307
16.2K
debug
debug@0xdbug·
Interesting, looking forward to this work. First step in hacking is always “nothing works magically and there are things lurking under abstractions”. Last step always is “breaking and exploiting it like a magic”.
zhiniang peng@edwardzpeng

Our talk: <Token Injection: Crashing LLM Inference With Special Tokens> will be present at #Blackhat_EU_2025 #token-injection-crashing-llm-inference-with-special-tokens-48830" target="_blank" rel="nofollow noopener">blackhat.com/eu-25/briefing…

English
0
0
0
204
debug retweetledi
Pedro Domingos
Pedro Domingos@pmddomingos·
Matrix multiplication is now a multi-trillion dollar business.
English
244
950
9.2K
443.8K
debug
debug@0xdbug·
@aall86 @33y0re @AlanSguigna Why take PF when software knows kernel may touch user memory and would want to protect against such accesses. If kernel intently wanted to touch user memory, it could do STAC/CLAC. Am I missing something on why kernel wants to receive PF instead of GP on inadvertent violation?
English
0
0
1
153
Andrea Allievi
Andrea Allievi@aall86·
@33y0re @AlanSguigna SMAP is not in use because it conflicts with LASS, where the Intel engineers injected a #GP instead of a #PF :-( Can someone explain to me why?
English
3
0
6
775
Connor McGarr
Connor McGarr@33y0re·
Very awesome!! It looks like code to support SMAP is beginning to appear the latest Canary build!
Connor McGarr tweet mediaConnor McGarr tweet media
English
4
19
131
16.4K
debug retweetledi
John
John@ionleu·
just search OPENAI_API_KEY on github. thank me latter.
English
353
879
18.1K
1.8M
debug retweetledi
BURKOV
BURKOV@burkov·
Apple did more for AI than anyone else: they proved through peer-reviewed publications that LLMs are just neural networks and, as such, have all the limitations of other neural networks trained in a supervised way, which I and a few other voices tried to convey, but the noise from a bunch of AGI-feelers and their sycophants was too loud. Now, I hope, the scientists will return to do real science by studying LLMs as mathematicians study functions and not by talking to them as psychiatrists talk to sick people.
English
214
833
8.5K
717.8K
debug retweetledi
Mickaël Salaün
Mickaël Salaün@l0kod·
I just published the fifth #Landlock newsletter! 🤓 - new kernel features: IPC scoping and audit logs - kernel fixes - library and talk updates - new doc - new open source Landlock users - RHEL support @digikod.net/" target="_blank" rel="nofollow noopener">lore.kernel.org/landlock/20250…
English
0
11
20
3.4K

Keşfet

@risc_v @fosdem @aall86 @33y0re @AlanSguigna @elonmusk @BarackObama @taylorswift13