🇪🇨-🇺🇸 caminante no hay camino, se hace camino al andar. 我不是随便的人，随便的时候我不是人。
Joined July 2012
- Tweets6.1K
- Following3.5K
- Followers159
- Likes79.3K
stelloprint retweeted
Imagine spending twelve years in school without studying:
— Homer
— Plato
— Aristotle
— Cicero
— Augustine
— Dante
— Shakespeare
— Milton
For centuries, these authors formed an important part of Western education.
Today, many students finish school without reading a single complete work by any of them.
How can we expect people to understand their civilization if they have never read the books that helped shape it?
stelloprint retweeted
Introducing Arete-27B: Qwen3.8-27B, post-trained to reinforce Western values.
Chinese open models are currently the strongest available, but they ship with state-aligned behavior in the weights. We built an 803-prompt eval across seven value categories, including freedom of expression, democracy, and privacy, and measured how often a model adopts China’s party line.
After fine-tuning on roughly 53 million tokens, Arete-27B's propaganda rate fell from 22% to 1%, in line with the Western models we tested. Its Hermes Bench score shows no significant performance tradeoffs.
The values in an open model are one fine-tune away.
Our eval, weights, and training data are available now along with the full blog post: nousresearch.com/arete
stelloprint retweeted
i went to bed and asked claude to paint an autumn tree overnight.
at around 4am, it realised that it didn't have some shades -- so it stopped and titrated its own paint: diluting red, sienna and madder step by step on a second palette, then measuring each pocket from a photo.
woke up to this beautiful watercolor painting. 303 brush marks, took about 4.5 hours 🍂 (still working on the speed)
stelloprint retweeted
hot take: you don't need a software factory, and i don't love the term. it's just marketing
all you need is good software engineering and investing your tokens in the right places. the more deterministic the better (and token efficient)
stelloprint retweeted
Pandilleros escapando de los soldados salvadoreños y pidiendo AUXILIO a la @ONU_es; haitianos dando un grito de libertad y agradeciendo al Presidente @nayibbukele por enfrentar a esos criminales.
Si esto no estuviera grabado, NADIE lo creería.
¡QUE VIVA EL SALVADOR! Sumate al vivo 👇youtube.com/watch?v=LIbggCdM…
stelloprint retweeted
Sorry guys. Software engineering and programming are dead. You just need to have perfect specs, unchanging requirements, an agentic deployment framework for high frequency releasing, and engineers with good taste and judgement.
stelloprint retweeted
Here's some fuel for all the Europoor haters:
Germany just told us we'd have to fire one of our 3 founders to qualify for funding.
I was in talks with a German state-backed fund of fund managing billions, with one single mandate: invest in German funds to drive innovation and economic growth in the country.
They literally told me they like us a lot and are also impressed by our track record and performance so far, and even though not everyone in their team is a fan of our thesis, they believe in it.
But they passed for a whole different reason:
Because our all-male team of 3 people isn't "diverse enough" lol.
they literally told me: hire two more female partners (impossible on the economics of a fund our size), or let one of the three of us go to make room for a quota hire. do that, and they'd strongly consider investing.
I won't even get into how discriminatory this is toward men in the first place. nobody takes that seriously anyway, and honestly I don't care.
but what actually gets me is something bigger: as a German, I have to face the fact that the whole world continuing to laugh at Germany and overall Europe is completely justified.
a system that quietly kills the smallest, hungriest teams, the exact ones this money was created to support. and then everyone acts surprised that talent and capital keep leaving the country.
stelloprint retweeted
Hacked my own VPS last night with abliterated models. Took less than 10 minutes and $5 in inference, and got root level access to the box and all credentials.
Lots to process and share soon, but feeling deeply in touch with my eastern european roots this morning...
stelloprint retweeted
Introducing Inkling 2 Small. Scoring 31 on the AA Intelligence Index, it's a big jump over its predecessor.
Inkling 2 Small is available for fine-tuning and inference on Tinker today, with full weight release planned soon.
thinkingmachines.ai/news/ink…
stelloprint retweeted
As I said in January 2024, no one should be taking mRNA COVID-19 vaccines because of their terrible safety profile. But for those who had to take one, Florida’s public health data strongly implies that thousands of Floridians died as a result of receiving Pfizer instead of Moderna. The number of lives lost in the U.S. and internationally is staggering, based on this analysis. The culture of approving vaccines without adequate safety testing and assuming “innocent until proven guilty” with these products absolutely, positively needs to end.
“Twelve-month all-cause mortality after initial COVID-19 vaccination with Pfizer-BioNTech or mRNA-1273 among adults living in Florida: Statewide cohort study”
Now appears as peer-reviewed study in Global Epidemiology!
In closely matched cohorts of nearly 2M Moderna & Pfizer vaccinees in Florida, the latter (Pfizer) showed substantial higher all-cause, cardiac, non-COVID and COVID mortality post 1st mRNA COVID vaccine dose.
These results were robust to adjustments on pre-vaccination cardiac hospitalizations as well as SARS-CoV-2 infections.
This study adds to a growing evidence that the mRNA COVID vaccines of Moderna and Pfizer may have different non-specific effects.
Recent study “In Vivo Blood Kinetics and Transcript Integrity of Three mRNA–Lipid Nanoparticle Vaccines in Humans” @acsnano showed significant differences in systemic persistence of nano-lipids and spike mRNA between Moderna and Pfizer mRNA COVID vaccines!
@MaryanneDemasi @FLSurgeonGen
stelloprint retweeted
Being right years ahead of others isn’t nearly as satisfying as people think. Nobody apologizes for treating you like a crazy. They just come back years later, repeating what you told them as if it’s some groundbreaking revelation because they finally heard it from someone else.
stelloprint retweeted
Let me tell you a story about how unethical Anthropic is.
In April, the Anthropic team began heavily courting our engineering team to tell them corporate secrets, technical details, and future plans for @cfosilvia.
They pretended their interest was due to our exploding popularity and accelerating compute usage, but it quickly became apparent that Anthropic was fishing for details on how we built our product and why it was so popular.
This included the Anthropic team trying to wine-and-dine our engineers, including a infamous steak dinner where someone told our team "don't worry about it, this is on Dario."
Our team was not born yesterday, so we quickly realized what was happening and refused to share sensitive information and technical details with the Anthropic team.
As we suspected, Anthropic was trying to get information from us, so they could build a competing product. Some people on their team must have a guilty conscious because they messaged us a few days ago to give us "a heads-up" that they were launching a competing product.
Thankfully, the Silvia engineering team has been aggressively working to get off the frontier models for the last ~ 6 months. That effort has been highly successful, because our compute costs are ~ 97% lower when we use our own model & hardware, the latency is lower, and the accuracy of answers materially increases as well.
The sad thing for Anthropic is they have now released an inferior product and they had to sell their soul to do it.
The only thing worse than being a trillion-dollar company that can't beat a 10-person team is being a trillion-dollar company that no one trusts and everyone knows is unethical.
Serious investors continue to choose @cfosilvia over ChatGPT, Claude, and other frontier companies because these products are not optimized to solve the actual user problem.
Today, Anthropic launched a Claude wrapper that is masquerading as a personal finance product. Best of luck to them.
But I will take the Silvia misfits vs the Anthropic losers any day of the week.
if you want to support people trying to do things the right way, and help investors be more informed, then consider signing up for Silvia today: silvia.com
stelloprint retweeted
there are very legitimate companies flooding reddit with bots to shape what eventually ends up in chatgpt
not just to simply recommend their product but to counter any narratives they don't like
stelloprint retweeted
my spidey senses are telling me the labs are, as we speak, using unguardrailed versions of their unreleased models to systematically reverse-engineer ALL of the internet’s software, apps, and games, then feeding the reconstructed source code straight into their training pipelines while simultaneously building guardrails to ladder-pull the rest of us
stelloprint retweeted
> be AI company
> scrape the entire internet
> ignore licenses and copyright
> download millions of pirated books
> use them to train frontier models
> spend tens of millions buying physical books
> cut off their spines
> scan every page
> destroy the originals
> lecture everyone else about intellectual property
> we are the good guys
We added two sentences to Claude Code's prompt on Opus 5.5 to steer it away from recreating someone else's specific game because using Claude to infringe IP is against our usage policy.
Original games, genre games, and games built from your own designs and assets are meant to keep working.
We'll continue to iterate to ensure we're striking the right balance. In the meantime, if Claude turns down something that's yours, send it my way.
stelloprint retweeted
From a startup founder: “it’s now a red flag for me if I don’t see side projects on someone’s resume / application, because it signals lack of autonomy. Because it’s now so trivial to build anything on the side, vs before.”
stelloprint retweeted
BREAKING: Revolut moved ~2,000 engineers onto self-hosted GLM for coding and pushed Claude Code and Codex to the exception list.
Per David Abitbol, who leads AI ops there: 56 Nvidia B200s, a blended 30 cents per million tokens, roughly a tenth of what they paid Anthropic and OpenAI. Engineering lead time improved 70%.
Here's the part that matters: the gains showed up when tokens stopped being rationed. Engineers quit saving AI for the hard tickets and pointed it at everything.
Frontier models become the exception, open weights become the default. That's a pricing problem for every lab heading to IPO.
h/t @OliverMolander
Would you move your team off frontier APIs if the bill dropped 90%?
