This letter from OpenAI is a real WTF nerf fest. Yuck.

Before they even finish stacking the first thousand V3s, both starship and starlink will already be a generation ahead of the hardware that flew yesterday. The 1 Tbps / 60-sat slide is the floor, not the 2027 operating point. This was like raptor 1 but for Starlink.

I’m ready to #RTXPowersPlay

This is a very useful post. It helped me Upgrade a loop skill I use for complex tasks like writing bigger apps, making videos, refining copy, and more. Well worth studying.

Build things then teach how you did it if you really want to learn things.

Agents have no agency. Agents amplify agency. Agents do not replace agency. So far, it is still only us humans that wake up wanting.

Watch this lovely animation produced by @LCSlates What I appreciate as much as the animation itself is the creative idea to prompt that! So cool.

.@ClaudeDevs ran into this today. I have plenty of usage left. I'm at 56% on Fable. Weather I have or have not hit some limit, it's very unclear. Nothing in any GUI would indicate I'm at the limit that would cause this. I very rarely hit limits due to the way I use claude so this was new(s) to me.

I wonder, do people appreciate the raw speed at which discourse, mostly an X phenominon, and implementation into products is happening now? It's really incredible. I think the answer is still no, they do not appreciate or even know. The gap is wide.

Steal this SuperGrok (Grok for X Premium users) prompt. It scans X for the prior week's tech/AI and classifies it through @michael_chomsky's classifier.dev which uses Jev from Typesafe + other methods attempting to suss out signal vs. noise for business owner's and founder’s once a week (or however often you choose to run it really) This is possible because Michael's tool is available for free AND accessible by Grok via HTTP page fetch; which is fun. Easy to modify as well if you want to tune it to specific news, industry, etc. Be patient when you run it, it is using Jev via classifier and grok will be still doing work as well so give it time to complete. Here is the text version of the prompt or as an image attached. -- cut -- Please read the last week's news about technology and AI, then run each story through classifier.dev by fetching the public page to surface what is unusually worth paying attention to and not just noise. If nothing is, say that. Do not classify in your head. Do not POST JSON. Use page fetch only. STEP 1 Last week = the 7 days before today. Gather ~40 distinct tech/AI stories from news and X. Dedup. Rewrite each as a statement of what the text is about in the simplest possible language: the fundamental subject matter distilled to its essence. Keep the source URL for every story. You will need it for the brief. STEP 2 For each sentence, fetch this page: classifier.dev/?labels=unusu… Read the page. The printed label is the answer. Do not drop stories before you fetch. If more than a third come back "unusually worth attention", fetch again with the sentence prefixed by: Default to ordinary noise unless this is a rare qualitative shift. If page fetch fails, stop: "Could not fetch classifier.dev. Filter not run." STEP 3 — BRIEFING # This week in tech/AI 2–3 sentences. What actually changed this week, in plain English. If nothing unusual, say that here and skip the next section. ## Worth your attention Max 5 stories. If zero, omit. For each story use this block: **Headline in plain English** [Source name](https://full-url) What happened, who did it, and why a serious reader should care — 3 to 5 sentences. Enough to understand the story without opening the link. Include the concrete fact (a number, a product, a ban, a breakout, a bill). Then one line on why this is not ordinary weekly noise. ## Safe to ignore A short paragraph naming the 3 loudest noise clusters and what they were about. Then 3–6 leftover stories as markdown links, not recaps: - [Story](url) - [Story](url) Rules: - Every claim in "Worth your attention" needs a clickable markdown link - Prefer the original article or primary source, not a random aggregator - No raw classifier labels in the brief - No "in a landmark move" - If you only have an X post as the source, link that post -- cut --

When a useful form of intelligence becomes 50–400× cheaper and 20–200× faster, you stop batching and start embedding it. Jevons paradox at work with Jev.

Words are an abstract form of human thought that can be processed by another human mind; making thought manifest. We humans certainly do need words until telepathy is commonplace. Recent Neuralink demos do show some promise in human to machine communication without words. An interesting stack that may be emerging is complementary rather than replacement. A person can think an intention; a BCI can turn it into text, speech, or a structured command; a compact on-device model like Needle 3 can then execute the command without generating more prose. Humans talking to other humans will keep using words (or an equivalent discrete code) for a long time to come I suspect. This reminds me of how Jev is described to work and leads to something potentially interesting. Jev, from TypeSafe AI, is a hosted “System One” model that also refuses to write. You send it state plus typed questions; it returns typed answers with calibrated probabilities in one parallel pass. No tokens of prose. This gives the stack a middle layer: • Intention: thought, speech, or UI. • Optional local binder: Needle 3 maps messy language onto tools or a schema on-device. • Decision gate: Jev scores, routes, or vetoes against live state, with a confidence number you can threshold in code. • Action: the device or agent does the thing. • Words: only when another human needs them. Building things with that stack should be very interesting!

Grok 4.6 + Grok Build was a major step. This one promises to be exciting indeed!

This virtual Enigma machine reconstruction/simulation and deciphering of a previously undeciphered message is extremely cool! Thanks for sharing that @carterleffen mvueh-enigma-solved.carterl.…

That pos by @deedydas was entertaining. And yes, the worries are real. But, 2026 is the latest verse of a very old song. Let's take a little walk through history because every time has its challenges. 🕛️ Athens, late 5th–4th century BCE The amount of anxiety among citizens who still remember the old ways is off the charts. Every month there is a new sophist or general claiming to have rewritten how a man should live. I cannot keep up. Should I even try? Am I already obsolete? What does my craft or my hoplite training mean now that rhetoric and paid teachers can make any clever speaker look like a statesman? Is traditional virtue still a 30-year career? Is my only hope to found a school or a colony? At the very least I should attach myself to a rising house or a new faction, but which one? Everything looks either bloated with Persian gold or has no real foundation. I joined one. Is my share of the booty or the land grant actually going to last, or will this be another Syracuse disaster? The city is booming. How in the world am I supposed to find a respectable wife when the old families are marrying for money and the new men have no pedigree? Will the next shift in alliances or the next influx of metics let me stay a citizen in good standing? Will I ever be able to afford a proper house inside the walls while the biggest concentration of wealth the Greek world has ever seen is happening around me? I have status and some silver. No one really cares. I do not know what to do with it. Am I supposed to start a school? I found some peace in children. But I am worried about them. How does one raise a son when the old gods and the old songs no longer command automatic respect? What are the rules of this new order? Socrates himself put the core complaint in the Phaedrus: the new technology of writing would produce “the appearance of wisdom, not true wisdom” and make people forget how to remember for themselves. Europe, ~1450–1550 (printing press and Reformation) Every season there are five new books, pamphlets, or maps that claim to overturn what we thought we knew. I cannot read them all. Am I falling behind? What does a lifetime spent copying manuscripts or memorizing the old authorities mean when a press can multiply a text overnight? There is so much print slop! Is the scribe’s or the scholar’s path still a durable career? Is my only hope to become a printer or a reformer? I should at least join a new workshop or a new confession, but which one? Most look either overvalued by princes or have no moat against the next edition. I started a press. It's doing well. Is my share of the profits or the new ideas worth anything, or will this be another pile of unsold volumes? Cities are filling with new men. How am I supposed to make a proper marriage when lineage and land no longer decide everything? Will the next edict or the next war of religion let me stay where I am? Will I ever afford a decent house in a university town or a trading city while the greatest wealth transfer in centuries is happening around me? I have some books and some standing. No one really cares. What is the point? Am I supposed to print my own manifesto? I have children. How does one raise them when any apprentice can now read ideas that used to be locked in monasteries? What are the rules now? Scribes smashed presses. Church authorities threatened excommunication. People complained there were already too many books. 🕛️ England, ~1811 (power looms and factories) Every week there is another machine that does in a day what took a skilled man a week. I cannot keep up. Should I even try to learn the new frames? What does a lifetime of hand-weaving or framework-knitting mean when a mill owner can hire a child to tend a power loom? Is the artisan’s trade still a 30-year career? Is my only hope to become a manufacturer myself? I should at least take work in a mill, but which one? Most look either overcapitalized or have no protection once the next improvement arrives. The mill is running. Is my piece-rate or my tiny stake worth anything, or will this be another failed concern that leaves me worse off than the cottage? Towns are exploding. How am I supposed to find a wife and keep a household when wages are falling and the old village ties are gone? Will the next Combination Act or the next enclosure let me stay on this land? Will I ever afford a decent room in Manchester or Nottingham while the biggest wealth creation event the country has ever seen is happening in the counting houses? I have survived. I do not know what the point of the work is anymore. Am I supposed to start my own shop? I have children. How does one raise them when the factory clock, not the seasons, now rules the day? What are the rules of this new order? The Luddites were not primitive; they were skilled men watching their bargaining power and status disappear. 🕛️ United States and Europe, ~1930s–1970 Every year there is another electrical system, assembly line, radio network, or computer that changes the tempo of work. I cannot keep up. Am I already behind? What does a trade or a white-collar skill mean when “technological unemployment” is the phrase economists keep using? Is any profession still a 30-year career? Is my only hope to start a firm or move into management? I should at least join a growing company, but which one? Most look either inflated by speculation or have no lasting advantage. The company is doing well. Is my pension or my stock actually going to be there, or will this be another firm that disappears in the next downturn? Cities and then suburbs are filling. How am I supposed to find a partner and afford a life when the old neighborhood no longer exists? Will the next immigration wave or the next policy shift let me stay? Will I ever afford a house while the largest expansion of material wealth in history is happening around me? I have the job and the appliances. No one really cares. What is the purpose? Am I supposed to start something new? I have children. How does one raise them when the rate of change itself has become the problem? What are the rules now? Alvin Toffler later named the feeling “future shock”: too much change in too short a time. Keynes had already flagged technological unemployment in 1930. And, don't forget how annoyed Socrates was because of the wicked technology that was, writing things down. P.S. I used Grok to help write this but final edited it myself. And, that is a perfectly acceptable use of AI in my opinion to express an idea.

p(doom) = p(nuts) / 0

This should not shock anyone. It’s a powerful way to learn and it is not restricted to creating software. And hiring someone like this is just good business. There is not much they can’t figure out.

The new voice model GPT-Live-1 is amazing! It handles accents particularly well compared to prior models. Building apps with codex and Astra is really easy. I just build a really fun language learning app based on a conversation I had yesterday with my son about it is (and is not) taught in high school here in California. This app closes that gap completely for conversational practice on demand in a fun and engaging way. So cool.

I'm not sure this is necessary. A really great harness layered with appropriate context as skills and data becomes domain specific as a result of the context layers.

I've been meeting more people that tell me, "I taught myself to code." Then, they show me what they made. Great stuff! But, they didn't write any code nor likely review a single line. The meaning of "coding" is changing rapidly. I'm aware this is blasphemy to a the small percentage of the world that crafted code by hand historically or still do. Those talents are rare, will become rarer, and will remain highly valuable where needed.