
@FriggHome will turn homes into quantum-safe DePIN hubs, securing Web3 from your living room by 2026. đ Legion of @sidexyz
Joined August 2021
- Tweets3.9K
- Following723
- Followers1.9K
- Likes14.5K
Pinned Tweet
Make DeFI Great Again...
DeFi was the beating heart of the 2021 BullRun. It gave meaning to Web3, then lost itself in noise.
With FriggLiquid and @FriggHome, 'Make DeFi Great Again' isnât just a slogan, itâs about giving DeFi back its purpose.
Fast, Fair, Resilient...
Velocity matters. After a few of us devs explicitly pushed for public endpoint prompt caching on X, $NBIS Token Factory finally shipped it within months and the Open Router volume has gone instantly parabolic (scaling past 80B daily tokens just today). Get ready for $NBIS Token Factory to now take a massive chunk of public endpoint Open Weights Tokens processed from competitors like @FireworksAI_HQ & @togethercompute
Raw serving performance like Latency/TTFT and TBT or Tokens per Minute/Throughput was never the blocker for $NBIS Token Factory. Without public endpoints having input prompts with cached pricing user cost bottlenecks alongside Token Factory simply not being picked as a default routing option on routing platforms like Open Router, $NBIS Token Factory for sure had a massive impact to its true potential token volume as I have highlighted plenty of times before.
That single engineering friction point was holding back production grade volume from providers like Open router, Vercel AI Gateway and more. With $NBIS closing the feedback loop and delivering this as fast as they just did, the instant usage spike on Open Router to $NBIS Token Factory speaks for itself. Keep and eye on this chart as this will continue to balloon.
Software + Hardware Infrastructure execution doesn't get better than this. $NBIS
Cached tokens are now live on Nebius Token FactoryâĄ
Reusing context across requests?
Cached input can help reduce costs. See cached-token usage in API responses and billing, with pricing per model.
More info: docs.tokenfactory.nebius.comâŠ
Babacool retweeted
Des agents privés pour sa sécurité.
Des mutuelles privées pour sa santé.
Des complémentaires privées pour sa retraite.
Mais aussi :
Des profs privĂ©s tellement lâĂ©cole nâassure plus rien.
Des milices privĂ©es pour supplĂ©er lâarmĂ©e.
Et le tout malgré des impÎts plus lourds et écrasants que jamais !
La faillite est totale.
đŽ EN DIRECT
La préfecture de police de Paris invite les commerçants à recourir à des agents privés avant les prochaines manifestations
l.bfmtv.com/Y3FV
Babacool retweeted
Las personas razonables cambian de opiniĂłn con la evidencia.
Los socialistas cambian la evidencia para no cambiar de opiniĂłn
Babacool retweeted
Pour info Ă 6% chaque français paye 1500 euros par ans juste pour les intĂ©rĂȘts
Sans un centime vers le remboursement de la dette
Babacool retweeted
Il y a des dĂ©cisions quâon appelle "prudentes" uniquement parce quâelles ressemblent Ă ce que tout le monde aurait fait.
Babacool retweeted
Meta now owns the entire AI stack.
1. Chips: MTIA
2. Infrastructure: Meta Compute
3. Foundation Models: Muse Spark
4. Applications: Muse & Meta AI
Every layer matters.
The more layers you own, the bigger your advantage.
AI infrastructure is becoming as foundational as electricity was a century ago.
The question is whether it's built with communities or on top of them.
@JensenHuang and @GavinSBaker circled the same point this week: data centers can be a genuine win for the towns that host them, but only if the builder shows up first, listens before breaking ground, and keeps the promises made on day one.
In Independence, MO, that meant:
- We paid off $270K in local school lunch debt before we ever broke ground.
- More than 99% of our construction workforce is local union labor.
- We signed a workforce partnership with Metropolitan Community College.
- Our closed-loop cooling system uses about as much water each year as a single restaurant, less than a typical gas station.
We show up under our own name. We work in the open with the community. And the commitments we make on day one are the standard we hold ourselves to for the life of the facility.
More on how we build AI responsibly: linkedin.com/pulse/building-âŠ
Gavin, spot on.
AI is bringing manufacturing back to America and reindustrializing the nation after decades of offshoring.
AI is creating demand that drives investment in our aging power grid and sustainable energy, powered by market forces, not subsidies.
AI is creating construction and manufacturing jobs across energy plants, chip fabs and data centers.
AI is creating new companies and industries. $400 billion has been invested in AI startups in the past six months alone.
Builders must partner with communities to build in their hometowns, earn trust and create local benefits.
We have an opportunity to create lasting benefits for communities across America and help America lead the next industrial revolution.
Babacool retweeted
Devant le MEDEF, nous avons assistĂ© Ă la primaire des Ătatistes.
Le plus Ă droite veut supprimer 250 000 postes de fonctionnaires. On ne retournerait mĂȘme pas au chiffre de 2017, avant quâEmmanuel Macron fasse exploser les effectifs.
Marine Le Pen assume « ne pas avoir une approche comptable des retraites ». C'est rassurant quand on parle du premier poste de dĂ©penses publiques. Elle maintient sa retraite Ă 60 ans, et commence son propos par lâimportance « du rĂŽle de lâEtat dans lâĂ©conomie ». MĂȘme introduction chez Jean-Luc MĂ©lenchon.
Les anciens Premiers ministres d'Emmanuel Macron promettent la main sur le cĆur de dĂ©sserer le carcan dans lequel il nous ont mis.
Et la gauche trouve que le pays le plus taxĂ© du monde a besoin de plus d'impĂŽts. De plus de normes. De plus dâinterventions publiques. Et dâannuler ses dettes.
Vous savez ce que proposent les complices du Casse. Pour les véritables solutions à nos problÚmes, lisez Le casse du siÚcle, dÚs mercredi en librairie.
I keep coming back to the branding & DA of Hermes Agent and @NousResearch. The blue, the typography, the way it feels like a magazine.
We donât talk about it nearly enough...
In AI, thatâs rare. Very rare.
1
2
53
La concurrence est féroce pour @NanoCorpHQ et @polsia une vrai alternative vient de faire surface! Venez tester
AI can now run and grow AUTONOMOUS online business (and actually generate money while you sleep)
No need for software or the Claude API.
- Top-tier marketing (website, cold email, SEO, Meta ads, LinkedInâŠ)
- Make sellable digital products and software
- Do your marketing autonomously
- Collect and withdraw payments
- API and CLI available
You can give it your idea and come back 3 days later with leads and a company running on free credits at bizos.cc
1
14
Babacool retweeted
Today, Kimi K3 is live on Nebius Token Factory, and weâre proud to be an official Day 0 partner for @Kimi_Moonshot
From the outside, the result is simple: choose the model and call the API.
Getting there was not. Over the weekend, teams across @nebiustf worked on early access, inference engines, optimization, capacity, testing, billing, the console, documentation, cookbooks and the launch itself. There were late nights and plenty of last minute fixes.
watching the launch come together reinforced why weâre building Token Factory.
AI teams have usually had to choose between the convenience of a closed frontier API and the control (but also the operational burden) of open weights.
Kimi K3 shows that open-weight models have reached the frontier. Token Factory makes them easier to run without asking every team to become an inference infrastructure company.
One API, more model choice, path from testing to production.
Proud of everyone who made this happen, and grateful to Moonshot AI for trusting us with Day 0.
Give K3 the hard problem.
Kimi K3 is now available on Token Factory.
Weâre excited to announce that Nebius Token Factory is an official Day 0 partner for @Kimi_Moonshot's Kimi K3.
Kimi K3 is the first open-weight model to reach frontier-level performance, a major step forward for open models.
It is built for long-horizon coding, knowledge work and reasoning, with native vision and up to 1M tokens of context. Artificial Analysis scores it at 57 on its Intelligence Index, just two points behind GPT-5.6 Sol (max). That puts Kimi K3 at the top of the open-weight field and firmly among todayâs frontier models.
Developers can access K3 through Token Factoryâs OpenAI-compatible API and console today.
Give K3 the hard problem.
Build with Kimi K3: tokenfactory.nebius.com/?modâŠ
My prediction : Market volatility is running out of steam.
Why? Because Trump's unpredictability is now 100% priced in.
When doing the unexpected becomes the norm, the surprise effect dies.
32
Don't bet against Elon...
New CursorBench results just dropped and Grok 4.5 is the story.
#3 overall at 66.7%. Right behind Fable 5 Max at 70.5%.
Now look at the cost column.
Fable 5 Max: $17.32 per task
Grok 4.5 High: $1.51 per task
That is Fable level performance at roughly 1/10th the cost.
And it beats Fable 5 High and Opus 4.8 Max outright.
The intelligence war just became a price war.
1
19
You need to read this!!âŹïž
Karp vient de dire tout haut ce que la Valley refuse d'admettre. Il décrit une peur trÚs simple : les boßtes qui adoptent OpenAI ou Anthropic ont le sentiment de brûler du temps sur des tokens, de ne capter aucune valeur réelle, et de finir dépossédées de leur propre IP. Ce n'est pas de la parano. C'est la fin d'un mirage.
Le mirage, le voici : croire que les labs vont avaler tous les verticaux. Pourquoi tout le monde y croit ? Parce qu'ils ont 3 Ă 6 mois d'avance sur les modĂšles, et qu'Ă chaque nouveau palier un pan entier du SaaS existant se fait remettre en cause. Ăa crĂ©e une illusion d'omnipotence. Une illusion. Rien d'autre.
Parce que la vĂ©ritĂ©, c'est que le modĂšle tend vers la commoditĂ©. L'intelligence symbolique â texte, image, son, vidĂ©o, code, toutes les modalitĂ©s â converge vers un mĂȘme plafond. Tous les labs finiront au mĂȘme endroit : quasi-parfaits, quasi-indistinguables. Et quand tout le monde est excellent, plus personne n'est diffĂ©renciant.
La tech ne saute jamais ce chapitre. Quand une couche se commoditise, elle devient un problĂšme d'infrastructure. Capex, Ă©nergie, silicium, latence, distribution. Ce terrain n'appartient pas aux chercheurs. Il appartient aux hyperscalers : GCP, AWS, Azure â et demain les datacenters orbitaux de SpaceX.
Ce qui nous amĂšne Ă la seule question qui compte : oĂč se loge la marge ? Loi de conservation des profits â quand une couche devient gratuite, la valeur migre vers la couche adjacente rare. Ici, elle remonte. Vers l'application.
Et la couche applicative ne se gagne pas avec un meilleur modÚle. Elle se gagne par la composée de quatre forces : effet de réseau, produit supérieur, GTM, marque. Et surtout leur effet cumulé dans le temps. Aucune n'est un problÚme de recherche. Ce sont des problÚmes de builder. De distribution.
D'oĂč le pronostic, froid : OpenAI et Anthropic ne mourront pas d'un manque de talent. Ils mourront de leur position dans la stack â coincĂ©s sur une couche qui tend vers zĂ©ro marge, avec un capex qui tend vers l'infini. Ils finiront rachetĂ©s Ă la casse par les hyperscalers qui, eux, possĂšdent l'infra.
Et les spin-offs ? Les labs surfinancĂ©s par les VCs, des centaines de millions levĂ©es sur un slide ? MĂȘme issue, en pire. Parce qu'ils sont majoritairement dirigĂ©s par des chercheurs. Des gens brillants, mais pas des entrepreneurs nativement. Or ce qui arrive n'est pas une course Ă l'intelligence. C'est une guerre de distribution.
La leçon est ancienne, Thiel l'avait déjà écrite : on ne capte pas la valeur qu'on crée juste en la créant. Les labs créent une valeur immense. Ils n'en captureront presque rien. Les builders de la couche applicative, eux, captureront tout. Construisez en haut de la stack. Jamais au milieu.
1
23


































