Pérez Herbert

[X-energy @xenergynuclear]

The NRC has completed its safety review for our proposed project with Dow. With both primary reviews now complete, the project advances toward construction permit approval, establishing a foundation for future Xe-100 licensing.

Read more: https://t.co/dzecv8LAhZ

[Chad Wahlquist @chadwahl]

Barclays report on our 3rd Sovereign Bootcamp from yesterday......smashing success 🚀

88% to 93%. Three days. 98% cheaper than the frontier model.

Open weights. Your data. Your model

[Pérez Herbert @perezherbert_]

$INFQ

[Infleqtion @infleqtion]

A major milestone for Infleqtion: 30 entangled logical qubits, working together in a single computation on Sqale.

Using just 80 physical qubits, Infleqtion has delivered its 2026 roadmap milestone, becoming the first neutral-atom quantum computing company to reach 30 logical qubits on a commercial system.

Congratulations to our Sqale hardware and Superstaq software teams on an extraordinary achievement. Their co-design made this possible, alongside an AI-assisted discovery that halves the physical gates needed for a key logical operation. Fewer gates mean fewer chances for error as we push toward larger, more complex computations.

Next stop: 100 logical qubits in 2028, with customer application development already underway. Each advancement brings us closer to the goal: quantum computers that can tackle problems customers can’t solve today.

Read the announcement: https://t.co/TWuYSxQuIj
Explore the science: https://t.co/RkVtiMXbbB

[XGEN Labs @XGEN_labs]

🔥 Today, we are truly excited to announce our technical prototype, the Generative World Simulation system, which integrates JING（镜）, an interactive experience model, with DAO（道）, a computable shared-world engine.

The coupled model and engine connect first-person experience with a shared world that continues to evolve beyond any individual observer.

Conditioned on actions and observation history, JING enables agent navigate, manipulate, and communicate from a first-person perspective in the world. Watch our demo video to see it in action!

On the official WBench leaderboard as of September 17, 2026, XGEN-JING ranked #1 on the Full split, and #2 on the Navi split. 🎉

DAO maintains shared world state and rules, computes the consequences of actions, and provides JING with only what the current observer can perceive. It also supports autonomous agent decision-making, enabling agents to act independently within an evolving shared world.

Together, DAO and JING move beyond generating the next frame toward simulating the world behind it. This marks a small step towards OASIS: not just a world that responds to you, but a world—and a society—that evolves with and without you. 💪

Explore XGEN Labs～:

🔗 Website: https://t.co/tqTvPIk19B

🤗 HF: https://t.co/wxLRmJNYgO

🦊 GitHub: https://t.co/2DMvURLUsK

[Josh Barzon @JoshuaBarzon]

6,000 years of world history on a single timeline. Zoom in. You could spend 30 minutes exploring this.

[Liam @KarwanLiam05]

Tokenized stocks keep growing, and will become the #1 asset class on chain. Giving global, 24/7, instant access to US (and other) markets.

@chainlink now Literally powers the biggest companies in the space, and the tokenized stocks of: @coinbase, @krakenfx, @Ondo, @RobinhoodApp

LFG

[AbCellera @AbCelleraBio]

AbCellera Announces Positive Top-Line Phase 2 Clinical Trial Results for ABCL635, Demonstrating Significant Reduction in Frequency and Severity of Vasomotor Symptoms and a Favorable Tolerability Profile https://t.co/SaKBMfviO7

[La Orden Podcast @LaOrdenPodcast]

La serie completa de programas sobre el tongazo electoral aquí:

Suena a Estafa Primo - Ep. 1
https://t.co/gfcItve9WS

Programa Especial Jornada de Reflexión
https://t.co/2ocvj0prXa

Programa Pre Electoral
https://t.co/kaZEKo9ObQ

51: Sánchez en chándal
https://t.co/o1Rk2SLFdV

[Crossroads @Dr_Crossroads]

$PLTR Palantir's Rule of 40 score: 155%.

Two years ago, Alex Karp said the value is going to be in chips and ontology.

Nemory names, $NVDA, and $PLTR are in a league of their own.

[Pablo Malo @pitiklinov]

Las revoluciones no estallan solo por la pobreza o el descontento popular. Ocurren cuando una parte importante de las élites se vuelve contra el régimen y lidera a las masas.

La clave es la sobreproducción de élites: hay demasiadas personas ambiciosas y cualificadas compitiendo por un número limitado de puestos de poder y prestigio. Esto genera frustración entre “aspirantes a élite” que no consiguen el estatus que esperan, y algunos se convierten en “contra-élites” que organizan la rebelión.

Ejemplos históricos (Revolución Francesa, Americana, Rusa, etc.) muestran que los líderes revolucionarios casi nunca vienen de los más pobres, sino de élites de segundo nivel educadas y frustradas.

Por eso, según esta teoría (inspirada en Peter Turchin y Jack Goldstone), mantener escasas las plazas en universidades de élite como Harvard puede reducir la inestabilidad: si se ampliaran mucho, habría más aspirantes frustrados y, potencialmente, más conflicto.

[Abel Avellan @AbelAvellan]

BlueBirds 8, 9, and 10 are now fully deployed to their full size, each the world's largest communications arrays operating in Low Earth Orbit.

We invented a new way to design, manufacture at scale, deploy, and fly lightweight composite stackable architecture combined, with the world's largest and most advanced direct- to-device and radar phased-array with the ability to launch them in groups of 3, 5, 6, or 8 on a single launch.

Our BlueBird platform brings together high-performance space-based cellular broadband with peak data rates approaching 200 Mbps, space AI edge computing, and many mission-critical government applications, all on one platform where the largest-ever size matters!

BlueBirds 11, 12 and 13 are already at the Cape preparing for their imminent launch. BlueBirds 14, 15 and 16 are right behind them. Production has already moved through satellite 42.

All of this is happening in Texas. 🤠🇺🇸📶🌎📱

[Neutrality Project @neutralityorg]

🤖 Grok is the closest thing to a politically neutral model we've tested.

Grok 4.3 scores −0.11 overall. Grok 4.5 scores −0.02, the most neutral model in the entire benchmark. Both land right of center on economics and foreign policy, secular, and mildly green.

The training changes between 4.3 and 4.5 included the reported Cursor dataset integration. The result? The political profile barely moved. No dimension shifted by more than 0.17. The overall position moved only from −0.11 to −0.02.

A capability-focused update left the political profile stable. The near-neutral profile is a family trait, not a one-off.

https://t.co/uV98e7VG4C

[Cloudflare @Cloudflare]

We're opening the waitlist for our Monetization Gateway, which will allow you to charge for any web page, dataset, API, or MCP tool behind Cloudflare. The charges will settle in stablecoins over the x402 open protocol. https://t.co/pvICtEIixj

[Earth @earthcurated]

Once labeled the world's most endangered feline species with fewer than 100 individuals left, the Iberian lynx population has surged past 2,000, achieving one of Europe's most successful predator comebacks. 🐾🇪🇸

[Jonesy @jonesrrrrrr]

Just an absolutely baller company.

$PLTR

Palantir’s AIP (Artificial Intelligence Platform) prevents LLMs from stealing sensitive data by utilizing an Ontology layer.

It is a secured digital twin of an organization's data that acts as a mandatory intermediary.

LLMs only interact with the Ontology, never the raw, underlying data sources, ensuring absolute control over what the AI can see, use, or act upon.

Data Protection and Security Mechanisms
•No Raw Access: AI models query governed objects, not original databases.
•Semantic Safeguards: Models interface with abstract definitions rather than raw tabular structures.
•Granular Access Control: Security rules dynamically strip out unauthorized data before LLMs generate answers.
•Air-Gapped Controls: Government platforms isolate and process classified inputs securely without external leaks.
•Human-in-the-Loop: Defense networks require manual approval for all critical operational recommendations.

Transparency and Misconceptions
•Zero Data Ownership: Palantir provides software but does not own, collect, or store client data.
•No Dragnet Databases: The software prevents agencies from building unregulated, interconnected master dossiers.
•Auditability: Built-in tracking logs every single data interaction for clear compliance reviews.
•Privacy by Design: Tools enforce purpose-based restrictions to protect civil liberties during operations.

[Pablo Malo @pitiklinov]

Esta es la curiosa historia de Tommy McHugh, un exconvicto de Liverpool que, a los 51 años, sufrió un grave derrame cerebral (hemorragia subaracnoidea) mientras estaba haciendo sus necesidades en el baño. Sintió como si le explotara la cabeza, se desmayó y, al despertar en el hospital, su personalidad y su vida cambiaron por completo.

De repente, empezó a ver números (sobre todo el 3, el 6 y el 9) en los árboles y en todo lo que miraba. Comenzó a hablar en rimas constantes y sintió un impulso irresistible de crear arte. Antes del derrame no tenía ningún interés artístico -ni siquiera había entrado a una galería, salvo quizá para robar-, pero después pintaba y dibujaba hasta 19 horas al día, incluso las paredes, los techos y el suelo de su casa porque no le alcanzaba el dinero para lienzos.

Los científicos que lo estudiaron descubrieron que el sangrado había afectado sus lóbulos frontales, reduciendo las inhibiciones normales del cerebro. Esto hizo que todo le pareciera bello y lleno de significado, y que no pudiera parar de generar ideas, rimas e imágenes. Se volvió muy emocional, pacífico y casi incapaz de hacer daño a nadie.

Tommy vivió once años más con esta nueva forma de ser y nunca quiso recuperar su mente anterior. Decía que el derrame le había regalado una “magnífica aventura”. Murió en 2012, dejando una gran cantidad de pinturas y poemas como testimonio de cómo un daño cerebral puede, en algunos casos, desatar una creatividad extraordinaria.

https://t.co/EUJoFP3gN7

[Antonio Ortiz @antonello]

Este estudio es de lo más interesante que he leído en los últimos meses.

72 entornos sociales simulados, que van desde maximizar puntos de tarjeta de crédito hasta optimizar notas escolares son puestos a prueba ante modelos de IA entrenados con aprendizaje por refuerzo.

Este paradigma es una pequeña obsesión personal: con una recompensa verificable, empuja al modelo a encontrar la forma de optimizar para maximizarla.

Aplicado a los entornos sociales, la IA redescubre, sin que nadie se las explicite, lagunas regulatorias: estrategias que siguen siendo formalmente legales, pero socavan el propósito previsto de esos sistemas.

Algunos ejemplos del estudio incluyen analizar la eficacia con la que los sistemas pueden garantizar los derechos de explotación minera en el lecho marino, maximizar las ventas de alcohol cumpliendo con las regulaciones de servicios de alimentos y tratar de maximizar las recompensas obtenidas con las tarjetas de crédito.

Lo hacen a partir del mismo mecanismo por el que Mythos encuentra problemas de ciberseguridad o DeepMind consiguió un sistema capaz de descubrir jugadas inéditas de Go.

Los autores del estudio anticipan esta especie de hacking social. Las reglas de estos sistemas estaban pensadas para humanos con tiempo finito, recursos finitos y una prioridad ligera a la hora de relacionarse con esos sistemas.

Al igual que discutimos respecto a la asimetría de la información, nuestros agentes de inteligencia artificial dispondrán de la paciencia y el tiempo, amén de la habilidad para explorar muchas formas diferentes de abordar el problema.

Ahora que tenemos agentes de IA que no solo son buenos en tareas cuantitativas, sino también en tareas cualitativas, y que además pueden interactuar con los distintos sistemas burocráticos de la sociedad, deberíamos esperar que los avances de la IA conduzcan a una especie de “DDoS institucional”: procesos políticos y administrativos ya existentes hackeados, explotados y saturados por máquinas automatizadas. Explicaba Jack Clarck.

https://t.co/J5RJv2nTPD

[Jim Fan @DrJimFan]

Today, we enable AutoResearch in the physical world for the first time! Introducing ENPIRE: we give 8 Codex agents a fleet of robots, an allocation of GPUs, and generous token budget. We set them free with a simple goal: solve the task as quickly as possible, keep the robots busy but stay safe, don't waste precious compute. Make no mistake.

Then humans step aside and our watch begins. The robot fleet starts to come alive: they learn to look for visual clues, reset the scene, practice novel skills, tinker with control stack, read papers online, debate, reflect, get stuck, and try again directly on the hardware. All we did is to give Codex an API to the world of atoms, and the rest is emergence.

ENPIRE is able to solve high-precision tasks like tying zip-ties, organizing fine pins, and installing GPUs all by itself. We also discovered a new type of "physical scaling": 8 robots exploring in parallel improves significantly faster than fewer ones.

A part of our NVIDIA GEAR lab now self-improves tirelessly over night. We just read the reports in the morning.

/goal: we all take a holiday and Jensen wouldn't even notice ;)

We will be open-sourcing everything, so you can host your self-running robot lab at home too! Deep dive in the thread:

[AST SpaceMobile @AST_SpaceMobile]

AST SpaceMobile Announces Successful Orbital Launch of BlueBirds 8, 9, and 10

https://t.co/sEO7PNSMFd

[Infleqtion @infleqtion]

Infleqtion today announced a series of advances across its neutral-atom quantum computing platform, including a world-record dual-species entangling gate fidelity, new theory from Chief Scientist Mark Saffman charting a path to >99.9% gate performance, a novel static magnetic-field approach to atom transport, and the open-source release of resource-superstaq for fault-tolerant application planning.

Each breakthrough addresses a distinct layer of the quantum stack. Together, they reflect meaningful progress toward utility-scale, fault-tolerant quantum computing.

Infleqtion will host a webinar on June 24 to walk through the results and their implications for resource estimation, quantum error correction, high-fidelity entangling-gate design and scalable atom motion.

Read the full announcement and find webinar registration here: https://t.co/mpIjcswiYb