En Carles Sala

My kids, maths, software, AI, motorbikes and music.
Co-founder of @oassisai, @datacebo and @pythiac_com.
Former researcher @lab_dai, @MIT. Co-created @sdv_dev.

[En Carles Sala @encsala]

A punt de presentar @OassisAI per primer cop al públic general com a finalistes del programa Impulsa Startup 2026 de @CambraSabadell!

[En Carles Sala @encsala]

@BenjaminDEKR Didn't they do something similar with 3.0 Pro?

I recall upgrading my apps to use it and a few weeks later I had to quickly change to 3.1 because 3.0 was being sunsetted.

[En Carles Sala @encsala]

Yesterday I was asking here why @Google was falling behind so much on AI.

Well, search no longer: They are busy drawing butts.

I get it, I get it... priorities.

[Google en español @googleespanol]

En un chat, terminar una frase con un punto a veces se siente como una declaración de guerra. 😅

Para darle más calidez y expresividad a la pantalla, los emojis planos evolucionan a 3D con Noto 3D:

✨ Expresividad en 360°: Diseñados con más dimensión, textura y personalidad para tus conversaciones.

🎨 Detrás del diseño: Conoce los retos creativos para transformar íconos 2D en personajes tridimensionales completos.

👀 Spoiler: Sí, pensamos absolutamente en cada ángulo (incluso en los que no te esperabas). 🍑

Descubre el proceso de diseño detrás de esta evolución:
➡️ https://t.co/wkUkDU763x
Para español: En Chrome > Clic derecho > Traducir.

[En Carles Sala @encsala]

I've been heavily using several of the big frontier models over API for some time now.

And I have no exact numbers, but I have the sensation that +90% of the times I get an error on the provider side, it's Gemini.

What is going on with Google? Why are they falling behind so much?

[En Carles Sala @encsala]

@senyorfont Hòstia Arnau, quina currada! Molt ben jugat!

Suposo que ho tenies més o menys preparat abans de l'anunci, esperant a que arribés el dia, no? O ha sigut realment des de zero en hores?

[En Carles Sala @encsala]

No només som els últims a rebre les millores, sinó que ara també som els primers a rebre els inconvenients: https://t.co/OkSE5lRmMo

Gràcies Europa

[En Carles Sala @encsala]

No puc més amb les maleïdes alertes #esalert. Sou una colla d'incompetents, @nuriaparlon!! Ara hem d'espantar a tothom cada cop que hi ha previsió de pluja?

Quines ganes de perdre aquest govern de vista. Podries prendre exemple de @sanchezcastejon i convocar eleccions, @salvadorilla.

[En Carles Sala @encsala]

Reading so many comments today about people worried about AI models that feel "pain" and can be "tortured" is really, really concerning.

"Painful", I suppose, according to these same people. I hope a had a button to repeatedly push to make it stop.

[En Carles Sala @encsala]

@levelsio Sounds to me like you did not cut off an acoholic. You cut off a douchebag who happened to be an alcoholic.

[Arnau Font @senyorfont]

Dots, els nous assistents personals 24/7 d’OpenAI no estaran disponibles a la Unió Europea. ❌🇪🇺

És esgotador.

[En Carles Sala @encsala]

@lluis_llach És demencial com us heu carregat el que un dia va ser l'@assemblea

[Bernat Dedéu @BernatDedeu]

La política ha comprat els mitjans públics, i és així com ens acostumem a veure documentals orientats a fer creure que els líders (sic) ho són tot, sense experts, amb mirada espanyola. M’entristeix veure la nostra tele en aquest nivell tan baix

[DataCebo @datacebo]

We just introduced GRMs: a new class of models that captures a relational database inside a generative model. It is now possible to do this.

For the past decade, I have had one goal: make machine learning and AI — predictive or generative — work with complex enterprise relational databases. The challenges are unique.

They are complex, messy, and span hundreds of tables. Nothing like academic datasets.

There is rarely any public data to test your algorithms against, much less to build a foundation model on.

Enterprises have immense value locked in this data. Bringing AI to it is worth solving.

https://t.co/hK8LrdvtaV

[Kalyan Veeramachaneni @kveeramac]

Pretrained general-purpose foundation models for relational databases. Good for hype and good on paper. Impractical in reality, for one simple reason: there is no publicly available data to train them on.

One of their promises was solving feature engineering. But KumoRFM, one of the relational foundation models, only managed to match RDBLearn (@LinjieXu). RDBLearn uses a pipeline built on deep feature synthesis primitives: a decade old algorithm from @MIT.

RelArena paper: https://t.co/3RvMe8ZlTr
RDBLearn paper: https://t.co/rsYMnjIBrf

[En Carles Sala @encsala]

@emergenciescat @112 Es veu que als responsables d'@emergenciescat i la @nuriaparlon no els explicaven el conte del Pere i el llop de petits

[En Carles Sala @encsala]

De debò no teniu res millor a fer que enviar una *$%& alarma a les 11 de la nit quan la gent dorm?!

I per dir una cosa tant òbvia com que "plou i potser hi ha riuades"? És que no teniu perdó!

Quines ganes de que hi hagin eleccions i es faci neteja al govern...

[Protecció civil @emergenciescat]

⚠️ Alerta #INUNCAT

Protecció Civil enviarà un ES-Alert al Baix Llobregat, Garraf, Alt i Baix Penedès i Vallès Occidental per pluges torrencials fins a la 1 h.

➡️ Risc d’inundacions sobtades
➡️ Extremeu precaucions en la mobilitat
📞 En cas de risc, @112

[Joel Joan @JoelJoanJuve]

O sigui que un fatxa de Vox, PP o del PSOE que vol la desaparició de la meva nació i llengua és tan català com jo? Segur, presi? Em vols fer sentir feixista per dir una obvietat? Tu si que fas discurs d’odi. Fàstic.

[En Carles Sala @encsala]

I have the feeling that everyone in the AI bubble is scrambling around, lost, without rumb. All sizes, solopreneurs and big corporations alike.

I keep seeing announcements of all sorts of random things. Often with huge overlap, often exactly the same. And everyone presents them as unique and SOTA.

And most of the times they are seriously lacking. Yes, they may serve a specific niche, but presented as big solutions for big audiences they fall short for 99% of their target, so they will actually get nowhere.

And the real challenge then is: how does one stand out with an actually good product among all this noise?

It turns out that product development became 10x faster (not easier!), but with it distribution became 10x harder!

So, is it really easier now to launch things as everyone seems to imply, or did it rather become harder altogether?

[En Carles Sala @encsala]

Molt i molt, eh?

De debò que ja costa massa de creure que pugui existir tanta incompetència

[En Carles Sala @encsala]

Alhora, tenim a Hanlon i la seva navalla.

Sempre l'he tingut en alta consideració, però últimament cal esforçar-se molt per atribuir-ho tot al cantó moralment acceptable.