Luiz Santana

Building tools to help the German Health System
CTO & Co-Founder at Digitale Patientenhilfe
Former- Tech Lead @Google. @Disney

[Luiz Santana @aroundluiz]

USA is becoming more Chinese than China itself

[Theo - t3.gg @theo]

It’s kind of wild that the two best AI models ever made are being restricted by the country they were built in

[Luiz Santana @aroundluiz]

@creatine_cycle I heard it is like Fable, cause you don’t have one

[Luiz Santana @aroundluiz]

@WarrenInTheBuff Not enough data centers for how heavy this model is

It sucks the entire resource

[Luiz Santana @aroundluiz]

@mandolinaes the people are amazing!
that's the main Spain thing for longevity

and yes, the olive oil

[Luiz Santana @aroundluiz]

@levelsio feels like a Severance episode

[Luiz Santana @aroundluiz]

Maybe they should have Tony Robbins wearing as an ad

[Nathie @NathieVR]

Evan Spiegel showing off the new Specs AR glasses to the public for the first time.

[Luiz Santana @aroundluiz]

@creatine_cycle Who framemog who?

[Luiz Santana @aroundluiz]

@RBouschery @peer_rich more like jun-aug

[Luiz Santana @aroundluiz]

Sometimes it kinda does, indeed

[Earth @earthcurated]

A study reveals that living in Spain feels like being on a permanent vacation.

[Luiz Santana @aroundluiz]

@brunocroh Eu tb e aí não crio nada. Mas se eu for honesto, o negocio é só fazer conteúdo com uma consistência boa que os assuntos vão vindo.

Só que sendo ainda mais honesto, não faço isso pq não dou prioridade suficiente

[Luiz Santana @aroundluiz]

Employees are allowed to “dogfood” / test the device.
Device falls into the water… what is the employee supposed to do?

GG… let the scuba diver have it

[Randy Pitchford @DuvalMagic]

A friend of mine found this watch a few days ago ~underwater~ when he was scuba diving near the island of St. Martin. He noted that the reverse of the watch indicates that it is a Google Pixel 5, which has not yet been announced, let alone released. It seems to be fine. The face indicates an empty battery, but seems to have enough reserve power to display the correct time. He asked me to have a look because he seems to believe I am connected in the tech industry. Who can help me out? Who knows how to find the person this unreleased watch belong to and let me know how can we return it to them?

[Luiz Santana @aroundluiz]

@brunocroh se tiver com um budget com folga, recomendo o Dell UltraSharp U3425WE

[Luiz Santana @aroundluiz]

@DanielLockyer i thought this was a running account

[Luiz Santana @aroundluiz]

How slow is big tech? You might ask…

This feature was developed over 2 years ago, when I was still at google.

They are advertising just now

[Google en español @googleespanol]

🚨 YA PUEDES CAMBIAR TU CORREO DE GMAIL 🚨

Ahora es posible cambiar tu dirección de correo electrónico actual de @gmail.com por una nueva sin crear una cuenta desde cero.

Lo que debes saber de esta actualización:

✅ Tu información (fotos, archivos y suscripciones) se mantiene intacta.
✅ Tu correo anterior se convierte en una dirección alternativa para que sigas recibiendo todos tus mensajes.

⏳ Una vez hecho el cambio, no podrás crear una nueva dirección de Gmail por los siguientes 12 meses.

[Luiz Santana @aroundluiz]

@kaanuzdogan not drown in calm waters.. be able to move in the water

[Luiz Santana @aroundluiz]

How are there adults that don’t know how to swim?

[Luiz Santana @aroundluiz]

UPDATE: GitHub confirms that cybercrime group looked at your code and said it sucks

[Luiz Santana @aroundluiz]

@brunocroh people can't read more than 2 paragraphs text

[Luiz Santana @aroundluiz]

@VictorTaelin weren't you praising it recently?

https://t.co/MFmtKlSvcP

[Taelin @VictorTaelin]

GPT 5.5 is much smarter than I thought

Yesterday, I did one-shots, coding, benchmarks, and was disappointed. Today, I did it all again, except via the API, which is now available. Results changed completely:
→ one-shot prompts went from bad to very good
→ excellent coding outputs, on both pi and holefill
→ benchmarks jumped, and now GPT *dominates*

I don't know what happened, I suppose there is something wrong with my Codex. In any case, truth is this model is very smart. It obliterated my benchmark, which is crazy because some of these problems were meant not to be solved. I'll need much harder tasks.

I also fixed 2 bugs that affected some providers:
→ added a retry for lost connection
→ removed the timeout limit
DeepSeek and Kimi wanted to spend more than 1 hour on my prompts, so I let them. Their results are much better now. Kimi K2.6 almost reaches Sonnet 4.6, although much slower.

Also this shows my points from last post were wrong

Again: this is a new vibe-coded bench, I'm focused on other things, so expect bugs and don't over-read this!

GLM 5.1, Gemma, Grok are not updated yet.

[Luiz Santana @aroundluiz]

how is it so obvious when something is written by AI?

will it change in the future?