
Crecí https://nitter.cf/t.co/1YNyZhYvwx (Canva lo adquirió en $320M). Crecí @SoyGutenberg a 9M en FB. Invitado de OpenAI a probar GPT-3 en 2021. Escribo cada día en 👇
Joined June 2022
- Tweets1.7K
- Following2
- Followers9.9K
- Likes1.2K
Esteban Constante retweeted
Agradecemos a @RealConstante por compartir su experiencia y conocimientos durante la conferencia realizada en la @ipn_escasto.
Gracias por contribuir a la formación de nuestra comunidad estudiantil y por brindar un espacio de aprendizaje y reflexión.
#EmpresasTrabajandoXTi
Esteban Constante retweeted
you can make minigames inside chatgpt using intelligent UI
tower defense games, mahjong,
and even a "simtower" game
Esteban Constante retweeted
El psicólogo que ayudó a John F. Kennedy a deshacerse de su complejo de inferioridad
🧵👇
Esteban Constante retweeted
"Something's got a hold of me lately
No, I don’t know myself anymore
Feels like the walls are all closin' in
And the devil's knockin’ at my door"
Esteban Constante retweeted
ive been testing dots in the past weeks as part of dev early access community testing
i enjoyed being able to continue working while i was traveling, without needing to have my computer open
it feels alive
it is proactive
and with spaces, it just keeps track of what i have to do, while going forward with those items that are waiting
can't wait for it to be available for everyone else!
Announcing dots. Dots work 24/7 for you, learn from your feedback, have their own computer, browser and can be connected to over 4k apps in our ecosystem.
They’re powered by Astra our best model yet. Included in your Pro plan, without drawing down on any of your usage. You can even call a dot while it’s working.
We’re getting you started with your primary dot today and soon you’ll be able to create entire teams of them.
Esteban Constante retweeted
for years i've grown accounts
most of them over 100k followers
many of them were left inactive due to not having enough time/resources to keep them running
but now AI has gotten to a point where i can just point my agents to them and keep them running endlessly
most ppl also have a bunch of inactive assets they're sitting on, waiting for the right moment
well, now it's the right moment
make a list of all the stuff you have
tell your agents to help you run it
sky's the limit
Esteban Constante retweeted
i want an AI that:
- is fast (1000+ tok/s)
- is cheap ($200 plans that feels unlimited)
- has a great image model (like gpt-image)
- can understand videos (like gemini)
- has taste (like opus 5.5)
- can one-shot video games (like Astra)
- can do true pixel art (like retrodiffusion)
- can talk to it (like gpt voice)
- can create music (like suno)
- can create cheap good looking videos (like seedance 2.5)
- has great mobile support
- can develop locally or in the cloud
- can create voices (like gemini's latest model)
- has really good memory
- can use my computer (like codex but faster)
- can work for hours without baby sitting it (like codex)
- can be used in different harnesses (like gpt does)
im probably missing some more stuff, feel free to add it below
Esteban Constante retweeted
quite impressive that a bunch of overdeveloped monkeys are making an intelligence far superior to them
Esteban Constante retweeted
walking while talking to gpt voice is a quality of life upgrade, specially now that you can chat with stronger models
Esteban Constante retweeted
who said ChatGPT sucked for designing websites?
i tested 100 different website design styles using GPT-6 Luna
(link below)
Esteban Constante retweeted
i'm terminally online, doing ai stuff, learning ai stuff, talking with ppl doing ai all day long, and im overwhelmed by everything that's happening
the acceleration is barely starting and i can't keep up with this
normies have no freaking idea of what's coming, do they?
Esteban Constante retweeted
broskis @MachgenAI have free Minimax h3 turbo free gen if your account has $25+ balance
time to make some vids then
🤝 Paid partnership
Esteban Constante retweeted
added to Pixel32Bench
you can see the improvement from the previous models
Please welcome GPT-6 Sol and GPT-6 Luna to the GPT-6 universe.
GPT-6 Sol and Luna build on the advances behind GPT-6 Astra, bringing much of its strengths into faster and more affordable models to support work at scale.
We’ve also made caching and inference more efficient, and we’re passing the savings directly to you: 50% lower API prices for Sol and Luna compared with GPT‑5.6 promotional pricing.
Esteban Constante retweeted
i've been looking at why jev's launch spread so fast
"the founder helped create chatgpt" explains why people looked
there's a lot more happening after that first look
13 things i think made it work:
Esteban Constante retweeted
most ppl dont want to host it
most ppl dont want to finetune it
most ppl just want an endpoint and something that works out of the box
GLiNER2 can perform any task Jev can. It has been open source for over a year, so if you see a lower score on your benchmark just fine-tune it 👀
It also supports extraction tasks (NER, Relation Extraction, ...)
Fine-tuning guide: github.com/fastino-ai/GLiNER…


















