Ideas para el nombre del grupo:
• La PPolla Records
• La Oreja de Van Vox
• Vito y los vitipaldis
• Efecto Caudillo
• El cuarteto de Vox
• Red Hot Chilli PPrs
• El canto del polo
• Jarabe de payo
• Arde Ceuta
• Viva Suiza
• Privatizadores del verso
DeepSeek engineer just published an essay explaining, in detail, why he's personally accelerating the AI that will make his own life's work obsolete.
His name is Shengyu Liu. His code sits inside DeepSeek V4.1's main attention kernel. A year ago AI could help him look up docs and hunt bugs. Now it reads CUDA, PTX and SASS on its own, analyzes stalls per instruction with professional tools, and iterates. His estimate: six to twelve months until it matches or surpasses him.
The part that got to him isn't unemployment. He thinks he'll stay in the industry. But the work shifts from writing kernels to piloting an agent that writes kernels. He used to sit at his computer for an afternoon meticulously tuning performance like a speedrunner chasing his own record. Those afternoons are ending. His phrasing: leave his talent behind in yesterday and become a pilot for an agent.
Then the essay pivots. Liu argues society ends up in one of two places, communism or Cyberpunk 2077. Either frontier AI becomes cheap infrastructure everyone can access, or a handful of companies own it and lock everyone else out. He describes the lock-out as a dead loop. You need the strongest AI to gain resources, but you need resources first to access the strongest AI. He says outright he doesn't believe Anthropic or OpenAI will voluntarily open their frontier. That's why he's staying at DeepSeek. To open-source enough to pull the world a step away from the second scenario.
MY THOUGHTS
This is the most honest essay I've read this this recently (sorry little angel).
Everyone warns about AI taking jobs. Liu is a practitioner at the frontier, personally speeding up his own obsolescence, who has done the math and is doing it anyway on ideological grounds. That's a different essay than the ones coming out of American labs.
The dead loop is the sharpest framing of AI inequality I've seen from inside the industry. You need top-tier AI to earn the resources that pay for top-tier AI, or you don't get in. That's not a warning about the future. That's a description of how enterprise AI pricing already works. Microsoft canceling its Claude Code licenses last week was the same story. Copilot users burning through credits in two hours was the same story. The people with the most money buy the best models, which earn them more money, which buys the next tier. Everyone else runs on the weakened version.
Anthropic and OpenAI dance around the concentration question in every safety essay they publish. Liu names it. His answer, that open weights are the only thing standing between us and a permanent AI class system, is what's actually motivating a lot of Chinese AI labor right now. Not nationalism. A bet on who owns intelligence in ten years.
He is a person building the thing that eliminates him, doing it on purpose because he thinks the alternative is worse. That's not a career anxiety post. That's a worldview. And the person writing it works on the kernel.
astonishing blogpost from Shengyu Liu (刘胜与, also known as interestingLSY/intlsy), kernel engineer at DeepSeek. The first part is his personal struggle with the fact that his work is about to render his beloved craft obsolete. The second is…
well. let's just say we agree.
Mucho análisis sobre Ceuta y Marruecos pero poco mirar un mapa.
El estrecho de Bab el-Mandeb conecta el mar Rojo con el océano Índico.
El estrecho de Ormuz, el Golfo Pérsico con el mar Arábigo.
Israel-USA han perdido ambos. Solo les queda controlar el Mediterráneo.
Hilo🧵
@Nanisimo@azukeka1917@Francis71827286 Clase media es un concepto sociológico/cultural, pero lo que de verdad define tu clase es en relación al capital, y ahí amigo, la "clase media" es clase trabajadora aquí y en conchinCHINA
Así es la esclavitud en las minas en el este del Congo, de donde sale más del 70% del cobalto del mundo, miles de esclavos diariamente extraen el mineral por apenas 2$ al dia para llenar los bolsillos a las multinacionales capitalistas.
El capitalismo que no te enseñan, así es como se sostiene el nivel de vida y de consumo en Occidente, en estas minas al menos hay 40.000 niños esclavizados que pican piedra para que Apple saque 4 modelos de Iphone cada año.
MIT's Nobel Prize-winning economist proved that AI is mathematically guaranteed to destroy human knowledge.
They published a massive NBER paper modeling the long-term impact of AI on human cognition.
And they found the most alarming conclusion in the AI literature so far.
It’s called "Knowledge Collapse."
Here is how human progress actually works.
When you struggle to solve a complex problem, you generate two things:
General knowledge about how the world works, and context-specific knowledge about your exact problem.
Normally, humans acquire both at the same time. You do the hard work to solve your specific problem, and in the process, you learn a general principle.
You share that principle. That is how human knowledge grows.
Then comes Agentic AI.
AI is incredibly good at giving you the exact, context-specific answer you need right now. It hands the solution to you on a silver platter.
So you stop doing the hard work.
And because you stop doing the work, you stop generating the "general knowledge" that society relies on.
Acemoglu calls it the "knowledge-collapse equilibrium."
When AI reaches a certain accuracy threshold, the incentive for humans to learn drops to zero.
Nobody verifies. Nobody explores. Nobody discovers new fundamental truths.
Society gets increasingly sophisticated automated outputs, while our actual capacity to generate new knowledge quietly erodes.
But here is the most terrifying finding in the paper.
Welfare is "non-monotone" to AI accuracy.
That means as AI gets more accurate, society actually gets worse off.