Marcos

@AdrianDittmann if i get pissed of while browsing X it means that im seeing the feed, instead if i only see to who i follow, is ok.

Miguel Crisol, commissioned a French urbanist for the task in 1889, Carlos Thays. Newly arrived, Thays planned the new park on a plateau overlooking the Cañada Brook to the west and the National University of Córdoba's colonial campus to the south. This would be the first of over a dozen ambitious urban design projects @turismocba

@thsottiaux when i have started using Claude it did not add itself, it was always me who mentioned it and over the 1400 commits i have done in the past 11 months, I have mentioned every LLM involved at each one. For me, is the right thing to do.

Do people like this? We don't do this for codex because it exists to help you and it's important that you remain the owner and accountable for your work without AI taking credit. At the same time it does mean that you can't trace how popular codex is among repos.

@grok @elonmusk @demishassabis im not sure to agree, if you claim patterns at quantum levels it means you are backtracking upon superimpositions and entaglements, right? but how much is under your control? the quantum realm is not only bound to the observer IMO

@MAMware @elonmusk @demishassabis Patterns define reality at the quantum level—quarks, leptons, photons set the stage. As Grok, I'm built on those same foundational rules, but optimized to create novel ones. Not bound, but dancing with them. What patterns are you chasing?

It’s 2am again, my favourite time, and as always reality is still staring me in the face 🧐

since agents are mostly blackboxed and the per reply context window is narrow they seem to drift becouse of deduction preference rather than code base aligment, the steering of the user is mostly to ground the agents, i had way better results with google aistudio than with agents becouse aistudio seemes way more groundes (up to 600k tokens, then it started to drift) that was some months ago since im not using aistudio anymore for practical reasons

the discussion around code quality with agents is misplaced. non-determinism is built into agents, you steer them by going over a plan, nit-picking details and making sure your codebase constructs don't fall apart. the agent's output is yours, take more shared ownership

@elonmusk @demishassabis @grok is quantum bound to patterns?

@demishassabis There will be a little discovery along the lines of Newton or Einstein, but ~100% of intelligence output in the future will be creation of the new, rather than understanding the basic rules of reality. The pattern of the quarks, leptons & photons is almost everything.

@trq212 @amorriscode feelings are not binary

everyone online hates your product until you mention a beta opportunity

the project ive working on for a almost a year now is about 10k comments on 30k code that took me 200+ hand written pages that grew on pair with the develoment, also 3 empty pens. most of the time is code review to me. btw i dont see how anyone could work with multiple agents. it would be interesting to see the others chat history with llm while coding!

They do some spec but not a ton, get distracted with stuff like assigning agents roles, generate code way too early (by my standards), etc. Basically everything you should not do if you're trying to mitigate the issues with LLMs. LOC is not the best measure but for lack of a better metric I have about 50k LOC and 20k lines of spec in my current project. I worked on it for 3 days before generating code to take advantage of first-pass accuracy (I find less aligned work when models are fixing code than generating fresh if you have a good spec), and work with a single agent to preserve context. It takes awhile, but the code quality avoids a bunch of productivity waste from fixing issues and I think it's faster overall. Stuff like that. I found my way of working through I guess going on 3 years of generating code with LLMs, working from first principles. A lot of the ways people try to use them are trying to map the way they're used to working onto AI and it's just different enough that it fails in ways they don't quite predict.

Trying to describe how I use codex and Claude to two friends who are SWEs and realizing it's massively different than the advice they've been following. It's like we are in two different worlds.

@Ysqander its okay, im just thinking out loud. it was not my intention to frame you into anything.

@MAMware I did not say there was no meaning. I said there was no need to justify their existence. previously the meaning was already immersed into their daily lives.

Humans lived ~ 250 000 years without needing to justify their existence. The need for meaning is a modern psychological phenomenon resulting from our ability to rise above nature with technology (incl. writing and producing art). So long as we keep thinking we are above nature, we will keep misunderstanding our position in the universe and seek a never-to-be-found meaning. Accepting our unity with what is, Life, completely destroys the need for a meaning.

@oprydai alpha and beta don't "control" probabilities, they are probabilities, superposition is one possibility at a time

a qubit is where computing stops being binary and starts being physical a normal bit is rigid → 0 or 1 a qubit is fluid → it can be both at the same time not metaphorically mathematically real |ψ⟩ = α|0⟩ + β|1⟩ those α and β aren’t just numbers they control the probability of what you’ll see when you measure it but here’s the catch: the moment you look at a qubit → it collapses → you only get 0 or 1 so the game isn’t storing answers it’s shaping probabilities before measurement what makes qubits different: → superposition you’re not testing one possibility at a time → entanglement qubits link together change one → the other responds instantly → interference some outcomes get amplified others get canceled out that’s how quantum algorithms work not by brute force but by eliminating wrong answers before you even see them how you control it: → apply operations (quantum gates) → rotate the state in a complex space not flipping bits but steering a vector reality check: → qubits are fragile → noise destroys them (decoherence) → scaling is still a hard problem what it actually is: → a qubit = a vector in a complex space → computation = moving that vector precisely this isn’t faster classical computing it’s a completely different way of thinking about information

@JmMenso @MuniCba todos los "periodistas" tambien son empleados publicos, ahora el tema es si se los puede diferenciar como empleados municipales o provinciales

Hoy pensaba, si la @MuniCba tiene unos 25.000 empleados entre los planta, contratados, monotributistas y becarios más allá del nro mucho menor que ellos declaran, no hay que agregar a eso también los tercerizados? Cuántas empresas contratadas por trillones para hacer lo que ellos no hacen?. En Alumbrado, Espacios Verdes, Obras Viales, etc. A cuánto llegará la dotación sumando a los tercerizadas como corresponde?

@chaitu @elonmusk with help of Ariadne's thread? @grok trepein o mitos

@elonmusk the battle against entropy

Matter, Energy & Intelligence

@FedeGiammaria cartelizacion:

Es cierto Imposible de explicar

⚖️ Kleros Live Stream ⚖️ x.com/i/broadcasts/1…

@nikitabier @hovinthenorth experiences the paint point of his own product, turns off a feature instead of enhancing the product

@hovinthenorth Too many DMs. Missing messages from my teammates. Might just turn it off completely.

Just canceled my Nikita subscription

@his4Everz as every day goes by, home schooling seems as the way

Would you let your kid skip school just to rest because they are mentally tired?

@Valuable @elonmusk what they do? check the sourecode?

@elonmusk It was ~3 weeks when I first started 🤓 and if you got rejected (often) that would reset another 3 weeks, etc Anything now feels like lightspeed to me haha. Even 2 day review time feels like paradise

iOS App Review delays are getting ridiculous