Alvaro Mateos

class Dev():
def __init__(s):
s.xps = ⟦⟧

def learn(s, xp):
s.xps.append(xp)

def create(s):
random.shuffle(s.xps)

def share(s):
print(s.xps)

[Alvaro Mateos @devprojects]

What if an AI could write an emulator from scratch?
zx-generation is a full JavaScript ZX Spectrum emulator — written entirely with Claude Code, Anthropic’s AI.
No human coding. Just prompting and iteration.

https://t.co/nc3OreAay3

[Alvaro Mateos @devprojects]

@DotCSV Mañana también saldrá el Sol.

[Alvaro Mateos @devprojects]

@claudeai Repost if you want a Fable reset. 😀

[Alvaro Mateos @devprojects]

@minchoi

[Alvaro Mateos @devprojects]

Adding new dimensions with Nano Banana 2 Lite

[Alvaro Mateos @devprojects]

@sama How long is the eclipse going to last?

[Alvaro Mateos @devprojects]

@thsottiaux Log in with different users

[Alvaro Mateos @devprojects]

@minchoi My two cents.

https://t.co/EjFCWBY3eu

[Alvaro Mateos @devprojects]

Already missing Claude Fable 5?

This flight simulator was built with a single prompt before it vanished.

Now it’s basically a relic :)

[Alvaro Mateos @devprojects]

El presente no es más que una iteración del pasado.

[Computer ♥ Records @ComputerLove_]

26 years ago.

[Alvaro Mateos @devprojects]

Already missing Claude Fable 5?

This flight simulator was built with a single prompt before it vanished.

Now it’s basically a relic :)

[Alvaro Mateos @devprojects]

@mundospectrum Y podeis jugar con ellos aqui 🙂

https://t.co/4QfO920WPw

[Alvaro Mateos @devprojects]

Lo de restringir modelos frontera no es una noticia más de IA. Si una empresa americana puede usar modelos que una europea no puede tocar, la ventaja competitiva puede ser brutal.

No es solo IA. Es productividad, ciberseguridad, ciencia, software, negocios...

[Carlos Santana @DotCSV]

💥 USA CORTA EL GRIFO DE FABLE 5 💥

Muy muy loco esto. Y un precedente muy grave que, ¿por qué no sería aplicable al resto de modelos?

[Alvaro Mateos @devprojects]

@DotCSV Me ha pillado bebiendo agua y menos mal que el tapon estaba pegado a la botella porque si no se me cae al suelo del susto.

[ElMundodelSpectrum @mundospectrum]

🟢 Álvaro Mateos, la mente tras Rocky 🥊, West Bank 🤠 y Capitán Sevilla. 🦸
📄 No te pierdas este extenso y apasionante artículo sobre uno de los autores más importantes de la época.
https://t.co/7S7ARGCtIP

#spectrum #zxspectrum #amstrad #commodore #msx

[Coollage Studio @coollage_studio]

1/ Claude Fable 5 just launched, and I've been looking at what people are actually building with it.

There's a lot of hype.

But a few examples are genuinely worth paying attention to.

Here are 7:

[Alvaro Mateos @devprojects]

Codex go remote:

https://t.co/y3A4LcCZUz

[Alvaro Mateos @devprojects]

@DotCSV Quien nos iba a decir que Mazinger Z se haría realidad.

[Alvaro Mateos @devprojects]

Pide a tu modelo que te genere HTML en vez de Markdown. Te ayudará mejor a entender lo que ha hecho. Funciona!

[Andrej Karpathy @karpathy]

This works really well btw, at the end of your query ask your LLM to "structure your response as HTML", then view the generated file in your browser. I've also had some success asking the LLM to present its output as slideshows, etc.

More generally, imo audio is the human-preferred input to AIs but vision (images/animations/video) is the preferred output from them. Around a ~third of our brains are a massively parallel processor dedicated to vision, it is the 10-lane superhighway of information into brain. As AI improves, I think we'll see a progression that takes advantage:

1) raw text (hard/effortful to read)
2) markdown (bold, italic, headings, tables, a bit easier on the eyes) <-- current default
3) HTML (still procedural with underlying code, but a lot more flexibility on the graphics, layout, even interactivity) <-- early but forming new good default
...4,5,6,...
n) interactive neural videos/simulations

Imo the extrapolation (though the technology doesn't exist just yet) ends in some kind of interactive videos generated directly by a diffusion neural net. Many open questions as to how exact/procedural "Software 1.0" artifacts (e.g. interactive simulations) may be woven together with neural artifacts (diffusion grids), but generally something in the direction of the recently viral https://t.co/z21CP5iQfu

There are also improvements necessary and pending at the input. Audio nor text nor video alone are not enough, e.g. I feel a need to point/gesture to things on the screen, similar to all the things you would do with a person physically next to you and your computer screen.

TLDR The input/output mind meld between humans and AIs is ongoing and there is a lot of work to do and significant progress to be made, way before jumping all the way into neuralink-esque BCIs and all that. For what's worth exploring at the current stage, hot tip try ask for HTML.

[Carlos Santana @DotCSV]

Esto es interesante: los modelos de Anthropic sufren más el "hablar" otros idiomas a nivel de uso de tokens.

En español el número de tokens usados se multiplica 1.62x respecto al modelo de OpenAI en inglés.

[Alvaro Mateos @devprojects]

GPT 5.5

[OpenAI @OpenAI]

Introducing GPT-5.5

A new class of intelligence for real work and powering agents, built to understand complex goals, use tools, check its work, and carry more tasks through to completion. It marks a new way of getting computer work done.

Now available in ChatGPT and Codex.

[Alvaro Mateos @devprojects]

@DotCSV No les convence el nombre. Mucho 5.