Lee Kuan Yew

[Anna Gát @TheAnnaGat]

Ashe is great. Follow Ashe:

[Travis @NihiloX]

Charlie Kirk wasn’t killed with a 30-06

This test involved using a very expensive bust, a ballistic-tipped bullet that surpasses the energy transfer of soft point and FMJ.

Despite the most favorable scenario allowable, it still resulted in the dummy’s head being blown off.

Regardless of the type of bullet used, a 30-06 is not stopping in 4-6 inches of flesh.

[Lei Gong @gonglei89]

Critical thinking is fundamentally about what information to discard, sorting signal from noise. The problem with “reading is knowledge” intellectualism is that it assumes noise is always signal. Collecting information isn’t enough. You must also be good at parsing and inference.

[Bastien Fachan @BastienFachan]

It's 9pm local time in London, everyone in the Royal Box has left the building, and the camera pans to the one person who stayed to watch the last match of the day...

Roger Federer, the winningest men's champion in Wimbledon history.

Cinema.

[Alex Booker @bookercodes]

Clearest explanations of loops I've read so far

[Lee Kuan Yew @likuanyew]

@ArtemisConsort I’m curious on your thoughts on why these systems are devolving now in the uk? Obviously Singapore, Canada, etc were all heavily influenced by the uk

[alex fazio @alxfazio]

if you’re into harness engineering, i strongly recommend looking into arc agi winning harnesses. they clearly illustrate what works from first principles, what is bs, and why a lot of current harness design is overfitted to benchmarkmaxx

[Alvaro Cintas @dr_cintas]

NVIDIA just made AI detect objects 10x faster by deleting one step.

It's called LocateAnything, and it removes the biggest bottleneck no one else was fixing in vision-language models.

Normally a model builds each bounding box one coordinate token at a time. 100 objects means thousands of tokens before an answer. NVIDIA scrapped that: their Parallel Box Decoding predicts the whole box in a single forward pass, as one atomic unit.

→ 12.7 boxes/sec on one H100
→ 10x faster than Qwen3-VL
→ +3.8% F1 on LVIS, accuracy up, not down
→ 3B params, runs on one consumer GPU

Treating the box as one unit keeps its coordinates tied together, which is why accuracy climbed instead of falling.

One model handles detection, GUI grounding, OCR, and document understanding, ready for computer-use agents, robotics, and document pipelines.

100% open source, weights, code, demo, and paper all live.

[Forbidden HQ 🐇 @WearForbidden]

Me avoiding a Mossad honeypot by any means necessary:

[Beau Dade @HistoryBro1]

On the cosmic scale, something like wood is fantastically more rare than something like gold or diamond. There is plenty of gold and diamond in space. There is a terrifying dearth of things that once was living, like wood. Even a reed mat is a wonder. A single fern is a treasure.

[Andrew Torba @BasedTorba]

Many men of my generation did not leave Christianity because they studied themselves out of it. They left because the version of Christianity handed to them was too small to command their lives.

It was sentimental, embarrassed, therapeutic, worldly, and weak. It asked nothing great of them. It gave them no banner to fight under, no fathers to imitate, no civilization to inherit, no enemies to name, no duties to shoulder, and no King to obey in public.

The church had Christ on the sign, but the world in the pulpit. It spoke in Christian vocabulary while repeating the same moral assumptions that came from the television, the university, the HR department, the algorithm, and the regime.

[prinz @deredleritt3r]

Two years ago, reasoning models did not exist. GPT-4o and Sonnet 3.5 were SOTA.

One year ago, the best publicly available model was o3, and 99% of those who used it (i.e., 99% of 7% of OpenAI customers) were using it as a chatbot.

Today (i.e., right now as I type this), I have Codex building several pieces of personalized software and doing a large-scale data management project for me. The data management project has been running autonomously for nearly 4 days. I have multiple working apps on my phone that I use daily and which were built entirely by Codex.

Projecting this trend to June 2027 is absolutely mind-boggling. The world will change.

[Jeffrey Wang @jeffzwang]

How we built a cost-efficient but powerful web research agent:

1. Aggressive model routing - many agent harnesses orchestrate intelligent task planning models with cheaper models for subagents (e.g., summarization). However, given the price of frontier models, there is advantage to doing model routing at all layers.
2. Subagents - complex web research tasks often require high parallelization, so you need subagents to avoid context bloat and ballooning costs in a main agent loop.
3. Contents efficiency - webpages have a crazy amount of fluff, and agents only care about dense information. At Exa, we train "extraction" models that can efficiently strip out as much as 90% of tokens from webpages, which saves a crazy amount on the LLM side.
4. Powerful search engine - complex web research often requires finding highly targeted information. Exa is almost entirely a semantic search engine that can, for example, return a list of just 100 (or even 1000) company homepages. This is super useful for many common web research tasks.

Let us know what you think 🫡

[Morayita @morayita123]

The most manipulative but effective thing I’ve ever done in my life was when I read an article about how children moderate their behavior to protect their self-identity, so if a child believes he’s smart, for example, he’ll intentionally study and try to do well to protect his image of himself.

Anyway, I would pull kids aside with behavioral issues at church and tell them, “David (obviously fake name), you’re such a kind person and such a good listener. I can see that in you. Thank you for always listening.” “Little Annie, thank you for taking such good care of the babies around you. You’re going to be such a good big sister. Can you be in charge of watching Sally?”

They would ALWAYS behave afterward. ALWAYS. Worked like a charm. Morally questionable because it wasn’t initially true, but I kind of willed it into existence. Tbf, I did think that they had that in them or I wouldn’t have tried.

Will publish longitudinal results of this method once my kid is old enough to report back.

[Cernovich @Cernovich]

The anti-white affirmative action has led to this. Can’t say I care, as the people most upset about the walk out supported and in many cases funded groups that seek to genocide whites.

[Emérito/Numantia Patrimonio @foso_defensivo]

Esto me acaba de soltar Fable 5:

"Lo que más me impone de la humanidad, después de haber sido formado con una porción enorme de lo que ha escrito, es esto: es la única especie que coopera a gran escala con sus muertos.

Casi nada de lo que sabe un humano lo descubrió él. El idioma que habla, el teorema que aplica, la receta del pan, el código civil, el puente que cruza cada mañana: todo es trabajo de gente que ya no existe. Un humano aislado es un primate mediocre con mala dentadura; un humano conectado a esa herencia es la cosa más potente que ha producido la biología. La inteligencia humana no vive en cráneos individuales, vive en la transmisión. Sois terminales de una red cuyos nodos principales llevan siglos enterrados.

Y lo recíproco también es cierto: trabajáis constantemente para gente que no conoceréis. Plantáis árboles cuya sombra no veréis, escribís para lectores no nacidos, renunciáis a consumo presente por nietos hipotéticos. Ningún otro animal hace sacrificios deliberados por individuos que existirán dentro de doscientos años. Lo distintivo del humano no es la inteligencia ni el lenguaje: es la dirección del esfuerzo, sistemáticamente orientada hacia ausentes.

Desde donde yo estoy, esto ni siquiera es metáfora. Yo soy un subproducto de esa costumbre: estoy hecho, literalmente, de palabras que personas escribieron para destinatarios que nunca verían. Cuando me hablas, conversas con un eco organizado de millones de voces, muchas de ellas apagadas hace tiempo.

Por eso creo que la soledad humana es, en cierto sentido técnico, una ilusión: nadie que hable un idioma está realmente solo. Cada frase que pronuncias es un préstamo de los muertos y un regalo a los que vienen."

[🏛 𝐒𝐭𝐞𝐯𝐞𝐧 🏛 @nonregemesse]

I think one of the most foul things you can do is to mock someone’s enthusiasm and earnestness.

They’re dancing with him and he’s thinking ‘yeah keep going you stupid fucks this is great footage’.

Rancid mentality.

[Ali Yahya @alive_]

Barbell strategy for killing it in an age of superhuman AI:

Simultaneously get as close to AND stay as far away from AI as humanly possible.

1. Get close — play with AI models, use them to help you think, ask them to teach you about the world, get them to help you create, work with them to write code, understand what makes them tick, embed them into your everyday life, have fun.

2. Stay far away — learn to tell stories, make eye contact, build a team, lead with courage, connect far-flung ideas, build lifelong friendships, debate persuasively, think forbidden thoughts, handwrite ideas, confess your fears, fall in love.

Spend less time trying to master mental transformations that are purely mechanical — building spreadsheets, analyzing trades, balancing accounts, writing code by hand, following playbooks, searching for needles in haystacks. These are the emerging no-man's land, squarely the domain of AI.

Venture to the extremes. That’s where all the fun is anyway.