Sergi retweeted
Oops! Missed the general strike in Spain today. Either way, good day to remind you that collective bargaining (strikes, unions, etc) is for losers, and that one of the most important and impactful things you can do in your life is to steer your professional career towards jobs with a high skill ceiling, so that you can negotiate with employers as an individual.
One of the reasons why the US is so ahead of Europe right now is because this simple fact is widely understood there, at a cultural level, whilst in the EU there are just millions and millions of people stuck with the mentality of a Victorian-era chimney boy. It's very weird to see so many folks with that relationship towards their work. Just completely surrendering control and assuming that any kind of improvement in their job and their lives can only come through external forces outside of their control.
It's a very defeatist attitude towards life, and funnily enough, it hasn't been _that_ true or effective in the past two centuries: the big wins in worker rights have come off the heels of great advancements in technology, and the same thing is going to happen with the advent of AI and automation.
But regardless of historical precedent, the fundamental issue is that this is just a very shitty way to live. If you can help it (and most people can!), why would you let your life's goals become hostages of the formless blob of the proletariat? What are the chances that the whims of the mob are actually what's most beneficial for you and your family?
I just don't see it man. I look at the people protesting in Spain and they're deeply miserable, and in the vast majority of cases this misery comes straight from the total abdication of their personal responsibility. In the 21st century, it's psychologically damaging to be willingly helpless towards external factors. You can't fix that by protesting. It's self-inflicted.
Anyway, back to work.
Sergi retweeted
UX latency is the most frustrating (lack of) advancement in computing over the last ~50 years.
IBM, way back in the 80s, found that user productivity shot up *exponentially* as system response time dropped. The "Doherty Threshold" was defined as ~400 milliseconds; the maximum amount of tolerable UX latency before interrupting user flow state.
It's unbelievable how many modern operating systems (*cough* win11) violate this threshold constantly.
Literally all you need to do as a Dev is pay attention. Add a "Doherty Threshold" to your test suite. There's so many different profiling tools, and especially with how good models are, really no excuse. Sure, some stalls are unavoidable, but that's when you pull out other UX tricks to at least give the illusion of responsiveness.
Crazy how much vintage computing hardware feels immeasurably more responsive despite having thousands of times less overall performance.
We built Control Room, a central system to track model workloads across our research infrastructure and stop them quickly when safety or security concerns arise.
It connects every research inference workload, including agentic sampling for model training and evaluation, data processing, and more. This works together with our misalignment monitoring systems, and halts model output within 35 seconds of a decision to intervene to stop or pause a job.
Control Room is just one part of our holistic approach for safe frontier model training and deployment.
openai.com/index/control-roo…
Sergi retweeted
actually, with AI I feel the push to become even more technical
there is so much to build, and I can go so much deeper with my explorations now
but in order to push the code to it's limits I need knowledge to know what's possible and what is correct
so I need to keep learning
HOLY: Anthropic's Claude Code engineer Addy Osmani is warning about a brutal AI paradox: the more coding you hand to agents, the more you risk losing the very skills needed to supervise them.
He calls it "cognitive surrender." AI-generated code looks pristine, tests pass, and engineers start approving work they no longer understand.
Anthropic's own research warns: "supervising Claude requires the very coding skills that may atrophy from AI overuse."
Osmani makes a striking observation: "When you're producing more than you can stand behind, you're no longer reviewing. You're hoping."
Sergi retweeted
We've unified plans across Grok, Cursor, and X Premium.
Introducing The X Subscription.
Sergi retweeted
launched a model
accidentally served trillions of tokens a day
29.4% of the fortune 500 showed up
raised a really big series A from @a16z
it's been 3 weeks
we would like to sleep now
OpenAI launched dots, always-on agents with their own computer, and a Decisions API on the same day. TypeSafe launched Jev, a model that only makes typed decisions. AI is splitting into coworkers and smart if-statements. Plus ClickHouse merges, AGENTS.md.
Sergi retweeted
One of the principles shaping how we build @rawtreedb is that every feature should be available through the API, UI, CLI and MCP following the same patterns consistently.
I recorded myself using @rawtreedb's MCP to create a Kafka connector and start ingesting.
Then I told Astra to make an edit for social media and... boom! 😅
Sergi retweeted
it's incredibly sad that clickhouse are so litigious that tinybird has to include this in their blog post. tinybird.co/blog/what-i-lear…
Sergi retweeted
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
Sergi retweeted
Using my app you can see that the electricity in Spain today is negative, meaning you actually get paid to use it
PD: it's called en2fe hours
Sergi retweeted
The magic of a sharded database is not having hundreds of servers. It’s making hundreds of servers behave like one.
planetscale.com/blog/making-…
Sergi retweeted
Beware people obsessed with outcomes instead of building outcome machines. Its worse than ever with AI, but these people existed before. Short term results above all else, etc. Don't fall into the trap. Invest in building strong fundamentals, invisible supports, and outcomes flow like water. An outcome machine.
Sergi retweeted
In Spain, my country, we have a mafia called @laliga. They block random @Cloudflare IPs whenever there is football on TV to "stop pirates". I don't watch football, I am trying to download a @HelloCivitai workflow to create videos for work, but I can't because my third world country blocks me. It's 2026. Shameful.
Sergi retweeted
Sobre derecho y ADE poco puedo decir pero en informáticas si quieres esas salidas excepcionales que te dicen que hay, tienes que hacer un esfuerzo exceptional también.
Nadie te va a dar un trabajo remoto de 100K en una empresa americana porque tengas un título.
Si lo único que haces es sacarte el título en España solo puedes aspirar a consultoras españolas donde vas a cobrar 20K.
Es más, el título es lo que menos influye cuando te mueves en este ámbito un poco fuera de serie.
En la startup donde trabajo el verano pasado contratamos un intern que con 16 años programaba compiladores en Rust, ni siquiera estaba en "ingeniería informática" sino telecos y no había terminado la carrera.
Este verano metimos a otro que nos hizo un SDK de Swift en nuestro repo open source, 19 años en primero de informática.
Hay gente que sale de 4° sin saber usar Git y se creen que están al mismo nivel.
Eso es lo que no te cuentan, que el título de por sí no vale una mierda.
Si quieres salidas excepcionales, demuestra que eres excepcional.
Si no pues las salidas son mediocres, al menos en España. – at Vigo, Spain
“Estudia algo útil, que tenga salida”.
Esa era la frase de siempre.
Ingeniería, Derecho, ADE, Informática, Sanidad.
Ahora hay miles de ingenieros trabajando de comerciales, abogados de administrativos, graduados en ADE de dependientes y informáticos de soporte técnico por 1.200-1.400 € al mes.
No es que las carreras “útiles” no tengan salida.
Es que la salida que hay no es la que nos contaron.
Empresas que piden máster, inglés, Excel avanzado, disponibilidad total y “capacidad de trabajar bajo presión”.
Y ofrecen 18.000-22.000 € brutos al año.
En ciudades donde un alquiler te come la mitad.
No es que la gente joven no quiera trabajar.
Es que el trabajo que hay no compensa el esfuerzo de formarse durante años.
¿Conocéis a alguien que esté trabajando “de lo suyo” y cobre lo que realmente merece?
HAR is live!
Run fleets of coding agents with deterministic validation gates, verifiable proof, and full observability. Extensible and customizable to your workflow and tooling.
github.com/os-factory/har
































