vivi

@crvvdev @ogisadaDMA Aren't those guys making like 200k+ at Riot? Other AC devs might be underpaid, I don't dispute that, but Riot, specifically, I find it unlikely. I also don't think comparing them to professional entertainers and entrepreneurs is fair, but that's neither here nor there.

@meridi6n @ogisadaDMA imo yes, they are being underpaid. streamers and pro players make much more revenue by doing nothing but playing the game, the anticheat business is stressful and the salary doesn't reflect how much the game companies profit from the game

Serious allegations regarding Riot’s Vanguard anti-cheat system continue to grow, including: • Claims that a Vanguard developer worked on a Vanguard spoofer called "Saturn" and other cheat related projects (a Javelin Anti-cheat Emulator) • Claims that another Vanguard developer also defended the first Vanguard developer to EA and said he was not part of Riot • Concerns that Vanguard may not fully delete user data after GDPR requests • A leaked Vanguard panel screenshot appeared online • Allegations that the first Vanguard developer mentioned also sent gifs in Discord servers containing signatures of a cheat blacklisted by Riot to ban legit people via Discord as Vanguard would think that the gif on your Discord is a cheat. • Claims that a third Vanguard developer leaked Vanguard blacklisted signatures and private information • Concerns that Vanguard can capture activity outside of League/Valorant when tabbing out often This post reflects my personal opinion and interpretation. It does not claim any verified facts, and should not be taken as a statement of proven truth.

@AnthropicAI "please give us subsidies + tax breaks or the world will end and we'll all become chinese slaves"

We've published a paper that explains our views on AI competition between the US and China. The US and democratic allies hold the lead in frontier AI today. Read more on what it’ll take to keep that lead: anthropic.com/research/2028-…

@crvvdev @ogisadaDMA don't think those guys are underpaid at all, probably just don't care & know nothing will come out of it

@ogisadaDMA devs usually underpaid compared to all the hassle AC development is; i'd not say it's a common thing that they cheat dev in parallel but I wouldn't be surprised if so lol.

@teortaxesTex It's so weird to me hearing him talk like this. Sounds almost depressed compared to his usual persona.

I think this is partially Trump's normal Middle Eastern diplomatic instincts, largely genuine admiration for strength and strongmen, but also, he probably appreciates a lot how Xi *never* moralizes or tut-tuts at him. Weak, cowardly Europeans berate him before submitting. Not Xi.

@VictorTaelin @LulaOficial tão bom quanto é o governo federal falando em construção de data center com o custo e matriz energética que temos... conversa pra boi dormir, mas sempre tem torcedor de time engolindo esses papos e ainda agradecendo a gentileza

@LulaOficial ta mas vc sabia que custa 1 milhão pra importar uma máquina de 500k e que isso torna impossível criar uma empresa de IA aqui imagina querer ter um país relevante onde suas startups precisam competir sem acesso a tecnologia básica kkkk

O Governo do Brasil zerou os impostos federais para compras internacionais até 50 dólares. O imposto de importação acabou. O Governo do Brasil está do lado do povo brasileiro. 🎥 Audiovisual/PR

@HyperDbg + SimpleSvm :p

Just another day at the HyperDbg project...

@vxunderground Will torrents ever come back?

Hello I have collected more malware. It's like, ... 200,000 malware, I think. I don't know. I've stopped counting. It is enough malware for your friends, family, extended family, neighbors, and co-workers. Please download it. The malware is lonely. vx-underground.org/Updates

Static Devirtualization of Themida/CodeVirtualizer. The techniques in this article apply to pretty much every virtual machine obfuscator with minor modifications. back.engineering/blog/09/05/202… Original Program & Devirtualized Output github.com/backengineerin…

@BackEngineerLab Thanks, another great one.

@teortaxesTex I agree that models will continue being released for the foreseeable future, but the point about hardware supply is a real risk as geopolitical tensions continue to increase. Hopefully China can provide an alternative to consumers medium/long term.

We've had Qwen 3.6 27B released since then Generally I think that dooming about open source stopping at some low level is asinine. This isn't 2024. We have 1T+ bases, we have agents, advanced RL&TTC methods. Even if new models stop coming, Q1 2026 frontier will be "open sourced".

@Tomas_Zubiri @gf_256 How so?

@gf_256 "responsible disclosure" You mean blackmail? The only responsible thing to do with vulns is not to disclose them. Everything else is racketeering.

responsible disclosure is dead🤦

@badlogicgames x.com/elonmusk/statu…

@badlogicgames Isn't there a Colossus 2 also? From the press release this deal applies to Colossus 1 only.

"we will use all compute capacity" did claude just kill grok? that's not very aligned (but also good riddance, grok)

@teortaxesTex Happy to be along for the ride!

🫡 moving the timeline's needle one anonymous shitpost at a time

@Peakanime0 "When I was a child, I spake as a child, I understood as a child, I thought as a child: but when I became a man, I put away childish things."

Completely mesmerized by this montage sequence from ghost in the shell... old school japanese background artists were in a completely different league 🔥

Left: Normal ASCII shader Right: Shape-aware ASCII shader ASCII characters are chosen based on their shape, so you get clearer edges Vectors are calculated and compared with WebGPU compute shaders It's based on an altered version of Alex Harri's approach. I'll link below

@VictorTaelin That is a pretty huge drop vs. 5.4. GLM and K2 also did very poorly; much worse than I would've expected.

Introducing LamBench . . . You asked me to make a benchmark, so I made it. It is a simple, old style Q&A consisting of 120 fresh λ-calculus programming questions. Some are easy, like "implement add for λ-encoded nats". Some are harder, like "derive a generic fold for arbitrary λ-encodings". It measures: - intelligence (% tasks completed) - elegance (BLC-length of solutions) - speed (completion time) Basically what I care about, other than long context. I made it today because I was excited about GPT 5.5. It didn't do too well ): (My first-day impression is that I can't tell the difference between GPT 5.5 and GPT 5.4. I would be lying if I said otherwise. I'd not be able to distinguish in a blind test. I need more time. It is much faster though.) This is a new, simple bench, so expect be bugs. Specially on OpenRouter models. I'll retest soon. Also, it was born saturated. V2 will be harder... ↓ Link and more charts below ↓

@yarden_shafir Initially I thought it was strange because in general you can't perform individual deallocations. But if the goal is to only ever protect a small, select number of data it makes perfect sense because it cuts down on bookkeeping associated with other allocators.

@meridi6n I think it's meant to have very limited usage. Why is the bump allocator strange in this context?

I checked and it's been 2 years since my last blog post??? So anyway, here's a quick blog post about KDP pool - the latest KDP feature that will replace the secure pool in future Windows versions: windows-internals.com/goodbye-secure…

@VictorTaelin ominous 👀

...