mx_twit

only here for bad memes and overhyped tech.

btw: Thinking is solved! trust me bro

[mx_twit @mx_twit]

same is true for software architecture and domain modeling

[Lorin Hochstein @norootcause]

I still find LLMs terrible at information presentation, even when they’re good at analysis. The reports they generate are not well-structured. The order that information is presented, and which details are relevant to the reader, these are things they are not good at yet.

[Dr_Gingerballs @Dr_Gingerballs]

Second law of thermodynamics. Statistical systems degenerate over time. Entropy increases. Anyone who thinks recursive self improvement is possible with these models is no different than people who think they can create perpetual motion machines.

[Garrett Wright @garrxttt]

Palmer Luckey says Silicon Valley’s elites have lost their connection to America: “They buy into political and social values they largely didn’t decide for themselves… and they’re insulated from the consequences.” 🇺🇸

“You don’t have to be that intelligent or articulate to understand the problem. A lot of Americans understand it. There are people who buy into a certain set of political and social values they largely didn’t decide for themselves, and they’re insulated from the consequences of their decisions.

When I was in Silicon Valley, there were a bunch of flags flying on the Facebook campus. A pirate flag, an Instagram flag, a gay pride flag, a world peace flag, and for a while, a Black Lives Matter fist. No American flag.

If you asked them, ‘Are you an American company?’ their official answer was that they were an international organization that transcended borders. They testified to Congress along those lines.

Culturally, it was true. They thought of themselves as part of this Davos elite. When you have a place where many people didn’t grow up in America with its values, you’re going to end up with a place that centers its concerns much more broadly outside the United States.”

[mx_twit @mx_twit]

@E_Boeminghaus steht IG eigentlich für Ideologische Gemeinschaft oder Interessenvertretung Grün ??

[Juliano Rizzo @julianor]

1/ AI is finding bugs faster than we can patch them, we hear "formally verify everything" more and more. So let’s look at the most verified kernel in existence, seL4, and what its latest release (Oct 30) fixed. 🧵

[H. Müller  🇩🇪 🎼 @HMueller_567]

@bitterlemmer @Axel_Bojanowski @welt 17 voll funktionsfähige und abbezahlte Kernkraftwerke, die leistungsfähigsten und sichersten der Welt, im Wert von 220 Milliarden Euro - einfach zerstört und gesprengt.

Aus ideologischen Gründen.

Wir werden von Wahnsinnigen regiert.

[mx_twit @mx_twit]

lol, rasputins heir apparently works at yandex.

[Лиса Ингерманландии @stars_latte]

На моей бывшей работе сменился ген.дир, и она сразу захотела, «шоб все было как у Яндекса» и наняла сотрудниц оттуда. Решила «оптимизировать процессы»

В итоге весь наш отдел (7 чел) к лету уволили :3 все это с улыбочками и формулировками в духе оттерапевтированных котиков

1/4

[Georgios Konstantopoulos @gakonst]

Legitimately can't believe people are still talking about managing fleets of sandboxes for their cloud agents infra. It's H2 2026. Harness out of sandbox. Virtualized bash & file system. Code mode. Sandbox only when you need to do something heavy. Not to edit files & simple data.

[Snicklink @snicklink]

Bill Gates got honeytrapped BEFORE it was cool 🫃

[mx_twit @mx_twit]

@TheCoinCollect8 will it improve on software architecture?

[Ian Macomber @iandmacomber]

Job families I'm bearish on: anything where the title + experience is vendor specific. Salesforce engineer, HubSpot developer, ServiceNow architect.

Vendors I'm bearish on: anything that's not agent-operable / agent-consumable by default, via any agent.

Businesses must design systems to be end-to-end agent-operable, and hire people who can contribute PRs across any part of the stack. If the rest of your org is getting 10-100x faster with agents, taking a dependency on a vendor that requires a specialist clicking around in a UI becomes the bottleneck.

Bullish on smart generalists, and companies who set up their codebases + cultures to allow anyone to gain confidence and build quickly in the best layer.

[Tendrils Compute @TendrilsCompute]

Tendrils has raised $30M led by @benchmark to rebuild general-purpose computing. @jaltma is joining our board.

At Tendrils, we believe that the future of general-purpose computing lies beyond legacy architectures for legacy software.

[Jarrad Hope λ @jarradhope_]

- AI holds economy hostage
- 65-75% of S&P returns in 2022 and 40-45% last 3 quarters, outstripping ALL consumer spending.
- 25x Enron-style debt hiding accounting.
- AI Hyperscaler Credit Default Swap markets 4x volume in 2 months
- AI firm circular relations.
- Macroeconomic Limits have been reached.
- Demand for Token Inference too small.
- Jensen Huang libido too large

[‎Wojak Codes @wojakcodes]

9/11 for all the electrical engineers who thought AI taking dev jobs was funny

[vittorio @IterIntellectus]

do you realise how fucking insane this is?!

[Nirav Patel @cmonkey]

Never mind, Fable 5.2 did it.

[KenFarmer ⒶⓋ @KenFarmer]

I don't think society has run out of smart people. I think we've made it financially and socially dangerous for them to speak honestly.

[Marco Mascorro @ CoRL 2026 @Mascobot]

Robotics will soon get bitter lesson'd by Claude, Astra, Gemini Argon, etc.

Just as custom coding models, embedding models, text-to-SQL models, computer-use models, etc., have been bitter-lesson'd over the past ~4 years, the same will happen to custom-trained robotics models, VLAs, world models, etc.

[Ramez Naam @ramez]

Anthropic is gaining exponentially less capability per billion dollars spent over time. It's now 8x as expensive to boost AI capabilities as it was 2 years ago, getting worse on a log scale. This includes METR report & internal models.

Credit to @alexolegimas, @sebkrier, and @michael_wiebe for inspiration.