Pinned Tweet
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
stay humble and stack sats
-----BEGIN PGP SIGNATURE-----
iHUEARYKAB0WIQRCpE7L4Z+8U2DOi92//mne3hoJzQUCaS+4RAAKCRC//mne3hoJ
zWHsAQD8OEofV7XLrUmsQju/blTP3wZLphRddwCLKn2UVvAb+QEA05ZCW7gcHLQk
DsYk9cWJXBxUhiXSK1PnPcTHy2lm6gc=
=Rb3o
-----END PGP SIGNATURE-----
Cindy retweeted
The number of users running nodes today is lower than it was in 2017!
Davidson Souza (@Erik17192799) takes the stage to present @getfloresta to the Japanese dev community! 🇧🇷🇯🇵
A deep dive into “SwiftSync”, and learning about Floresta on Mobile 🌲📲
@TBB_JP @btcinsider__
O Google DeepMind afirma ter provado matematicamente propriedades de segurança essenciais do código de máquina por trás do Signal. Não do projeto no papel, mas do código compilado.
Pushmeet Kohli, da DeepMind, diz que é a primeira vez que o código de produção de um protocolo amplamente usado é verificado formalmente "por qualquer método, com ou sem IA" e com essa profundidade.
• As provas cobrem o handshake pós-quântico PQXDH e o Triple Ratchet do Signal
• Propriedades provadas: sigilo futuro (forward secrecy), autenticação, segurança pós-comprometimento e sigilo do handshake
• Agentes de IA escreveram as provas em Lean
• Especialistas humanos revisaram cada especificação antes de qualquer prova ser executada
A verificação formal é extremamente complexa e demorada, e por isso quase nenhum software passa por ela. A DeepMind defende que a IA pode mudar isso.
Se isso se confirmar, o padrão para ferramentas de privacidade sobe. Provas em vez de promessas.
Google DeepMind says it has mathematically proven key security properties of the machine code behind Signal. Not the design on paper. The compiled code.
DeepMind's Pushmeet Kohli calls it the first time the production code of a widely deployed protocol has been formally verified, "by any method, with or without AI," and at this depth.
• The proofs cover Signal's post-quantum PQXDH handshake and the Triple Ratchet
• Properties proven: forward secrecy, authentication, post-compromise security and handshake secrecy
• AI agents wrote the Lean proofs
• Human experts reviewed every specification before any proof ran
Formal verification is extremely complex and time consuming, which is why most software never gets it. DeepMind's pitch is that AI can change that.
If that holds, the bar for privacy tools moves. Proof instead of promises.
suspension of mineral water sales due to cartel threat 🇧🇷
g1.globo.com/ce/ceara/notici…
10
438
Cindy retweeted
We were taught that Earth orbits the Sun in a closed circle. In reality, Earth has never returned to the same point in space once in 4.5 billion years.
Right now, the Sun is carrying Earth and every planet around the center of the Milky Way at 230 km/s (828,000 km/h), more than 18,000 kilometers every 80 seconds.
Because the Solar System is tilted 60.2° to the galactic plane and bobs up and down through the galactic disc roughly every 30 million years, the planets trace open 3D corkscrew helices across deep space rather than flat loops.
The last time the Sun stood where it stands today, one Galactic Year (230 million years) ago, the very first dinosaurs were only beginning to appear on Earth. Since forming 4.6 billion years ago, our Sun has completed only 20 laps.
Cindy retweeted
A Binance acaba de sofrer o MAIOR HACK da história das criptomoedas.🔴
US$ 6,3 bilhões em 10 blockchains, US$ 4,9 bi nos primeiros 14 minutos.
Mas nenhuma chave privada vazou.
Vou te explicar como foi, para onde foi o dinheiro e quem possivelmente está por trás.
Thread.👇🧵
Cindy retweeted
The Global Majority will lead this effort. Technology is no longer the bottleneck, our expectations are.
.@jack thinks bitcoin payment adoption isn’t where it should be. We want your ideas on how @Blocks, Spiral, and you can increase adoption within 5 years. Our trajectory needs changing. Here’s more on our thinking. Now we want yours: github.com/spiralxyz/bitcoin… View article
Cindy retweeted
Congrats, Renato! This is what our fellowships are for: helping developers find a path into Bitcoin open source and toward sustainable work, whether through a Vinteum grant or support elsewhere in the ecosystem. Proud to have supported your journey. Now let’s get 1.0 done.
Grateful to Spiral for backing rust-bitcoin. Thanks to @Vinteum_org for getting me here and for all my mentors. Let's get 1.0 done.
Cindy retweeted
Today at @tabconf, the Vinteum community at the Hackeroom.
1:00 PM
Monitoring the Bitcoin Network
@Octavio_lucca
1:15 PM and 3:00 PM
Two workshops: Bitcoin Setup Across Seven Domains and DuraShare seed backup
@renatoslopes
Check it out!
🤖 Made with AI
Cindy retweeted
Vinteum fellow @IsaqueFranklin0 presented "Limbs, carries and reduction: field elements in libsecp256k1" at @tabconf today.
The talk built a mental model of how libsecp256k1 represents 256-bit numbers as 5x52 field elements. Want to know more about it? The written deep dive behind it covers the same ground:
harlock.xyz/deep-dive-into-m…
Cindy retweeted
Karpathy AI funnel
🤖 Made with AI
Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.
- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.
The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.
So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.
One underrated aspect of multi-vendor multisig is that it improves the most important self-custody element: knowledge.
Different UI and interaction methods helps assimilate important concepts.
Comparing solutions yourself instead of being influenced is another factor.
these are the electronic voting machines
many brazilians believe they are the best voting system in the world. 100% secure.
and tomorrow is circus day in Brazil: elections
the paying audience will watch the polarized clown-candidates
whoever wins will benefit from the money printer and keep the population poorer
“Give me control of a nation’s money and I care not who makes its laws”
Eikos 27b foi avaliado pelo Akita! 🔥
Post novo: "Entendendo o Hype do Jev: pra que serve?"
akitaonrails.com/2026/10/03/…
Finalmente resolvi fazer um artigo que desce mais a fundo nessa história maior sobre classificadores estilo Jev, toco um pouco nas demais alternativas (fechadas e open source), e pra explicar a literatura e o funcionamento, desta vez tem uma novidade: este artigo foi escrito a 4 mãos, com colaboração do pesquisador da Lua Vision, Paulo Camara.
Acho que você não vai achar artigo mais explicativo sobre esse assunto! 🦾
Eikos está aberto (MIT).
Modelos de decisão (System One) de 4B e 27B: decidem numa única passada e dizem a probabilidade de cada opção.
• 27B: 82,9% no JevBench hard (Jev: 73,0)
• 2,4% de erro com ≥90% de certeza (Jev: 5,9%)
• 64k de contexto
Mesmos itens, mesma pontuação: 8 baterias contra o Jev e a Laya.
Pesos, dataset e código: huggingface.co/collections/c…
github.com/caiovicentino/eik…
Eikos, um modelo de decisão open source (MIT), alternativa ao Jev e ao Laya.
Sai em dois tamanhos, 4B e 27B, com versões quantizadas (FP8, INT4 e MLX para Mac).
O que muda:
• lê até 64k tokens (Laya: 1k, Jev: 32k)
• no JevBench, nível difícil: 82,9% (Jev: 73,0%, Laya: 35,1%)
• quando tem ≥90% de certeza, erra só 2,4% das vezes (Jev: 5,9%)
Eikos vem da filosofia de Aristóteles, "O provável".
Logo mais vou liberar no Hugging Face/GitHub pesos, códigos e datasets que criei sinteticamente.
my last few days have been spent between code and tears
my aunt passed away, and she was a huge ex of faith in Jesus to me
she used to say: "if what I want happens, it was Jesus, and if it didn't happen, it was still Jesus"
thank you for teaching me and for taking care of our family
your legacy lives on with us
our ancestors lives on us
12
62
1,278
































