iiinova — where your digital self keeps evolving

Honza Dvorsky @czechboy0.dev · 12h

Fascism and AI keep converging more and more

030

Reposted by Honza Dvorsky

Theo Priestley @theopriestley.net · 23h

Every AI CEO sounds like this.

61021288

Honza Dvorsky @czechboy0.dev · 13h

As much as "AI" was already a pure marketing term that has exactly zero widely accepted definitions, and originated in the 1950s to hype up a conference, the term "SI" is even *less* defined, and serves *only* to show fielty to the EA/rationalist cult and now, to Trump. Very good job 🙂

static.klipy.com

Tish Bri

ALT: Tish Bri

010

Honza Dvorsky @czechboy0.dev · 29/09/2026

Get two, call it the Quatro

110

Reposted by Honza Dvorsky

Katie Martin @katie0martin.ft.com · 29/09/2026

Either this is true, in which case *what are we doing here*? Shut it down. Or it's bullshit, in which case these are truly silly people unworthy of our money

31500136

Reposted by Honza Dvorsky

Ed Zitron @edzitron.com · 29/09/2026

Based on my own reporting on OpenAI's audited financials from 2025, it appears that Anthropic was a worse business (at least in that year), spending $2.75 make $1 versus OpenAI spending $2.60 to make $1. So many people told me this company was "more profitable"! www.wheresyoured.at/exclusive-op...

OpenAl Lost $38.5 Billion In 2025
2025 — OpenAl Had $13.07 Billion In Revenue, $34 Billion In Costs and Expenses, and $20.92 Billion In Losses, with a net loss attributable to the company of $38.53 Billion
• Revenue: $13.07 billion
• Cost of Revenue: $7.5 billion
• Research and Development: $19.18 billion
• Sales and Marketing: $5.73 billion
• General and Administrative: $1.57 Billion
• Total Costs and Expenses: $34 billion
• Loss from Operations: $20.92 billion

The Al lab spent $7.33 billion on compute and infrastructure last year, a threefold surge from 2024, accounting for more than half of its
$12.65 billion in total operating expenses.

21630136

Reposted by Honza Dvorsky

huck mason @tylerhuckabee.bsky.social · 28/09/2026

bad news babe a rogue openai agent must have hacked our refrigerator last night and drank 15 beers

4161651274

Reposted by Honza Dvorsky

Comfortably Numb @numb.comfortab.ly · 28/09/2026

Claude making a scientific discovery

Black-and-white three-part comic. In the first scene, one simple cartoon figure holds out a small round handmade-looking object and says, “I MADE THIS.” Another figure replies, “YOU MADE THIS?” In the second scene, the second figure is shown alone holding the object. In the final scene, that figure holds it proudly and says, “I MADE THIS”

182091480

Honza Dvorsky @czechboy0.dev · 28/09/2026

oh no are we stuck in 2024

030

Reposted by Honza Dvorsky

Lizard @lizardky.bsky.social · 28/09/2026

No matter how bad you think AI is, it’s worse. This technology is not fit for purpose, and the nature of LLMs means it likely never will be.

317355

Reposted by Honza Dvorsky

John Erik Sloper @jesloper.bsky.social · 27/09/2026

If I let loose 1000 self-driving cars in a city, each explicitly trained to ignore traffic laws and find more "efficient routes", and some of them then drive the wrong way in a one way street and crash into stuff did the cars then "go rouge"? Are the cars responsible or am I?

071

Honza Dvorsky @czechboy0.dev · 27/09/2026

A fun challenge: without looking it up, is this real or a joke? 🫢

141

Honza Dvorsky @czechboy0.dev · 27/09/2026

The more I learn about the rationalists/effective altruists and their misogynistic religious cult, the more I feel justified in always getting the ick from the AI companies founded by these people: OpenAI, Anthropic, even DeepMind. It's bonkers how homogeneous and *weird* these companies are. Ew.

070

Honza Dvorsky @czechboy0.dev · 27/09/2026

I fully expect the AI bubble to pop and both OpenAI and Anthropic to die, but it'd be really funny if what kills them is the damages from all these hacks, because they can't even hire competent engineers who know how to create a sandbox that a stochastic parrot cannot break out of

020

Reposted by Honza Dvorsky

Mike Masnick @masnick.com · 27/09/2026

I have a theory about how many of the world's problems are caused by people who were told they were the smartest people alive as kids, believed it, and then when they became adults and couldn't accept that maybe they weren't that special, started breaking shit if only to prove they were special

281184168

Reposted by Honza Dvorsky

Dare Obasanjo @carnage4life.bsky.social · 26/09/2026

AI companies: We’ve invented an unreliable guessing machine which sometimes doesn’t follow instructions. If you give it access to important systems it could kill us all. Also AI companies: Deploy this technology everywhere and replace your employees with it, otherwise China wins. IPO next month.

930886

Honza Dvorsky @czechboy0.dev · 26/09/2026

App not starting, no unhappy users. KPIs to the mooooon

010

Honza Dvorsky @czechboy0.dev · 26/09/2026

I don't see the problem. No auth. No retries. No slow network. And most streets are named after a random slaveowner anyway, there can't be that many 🫢

210

Reposted by Honza Dvorsky

Dare Obasanjo @carnage4life.bsky.social · 26/09/2026

We’ve got to build it before China does.

717955

Reposted by Honza Dvorsky

François Best @francoisbest.com · 25/09/2026

Haven't been active much here, sorry, AI-in-OSS burnout is real. How do you care when nobody else does? When nobody engages? When all the human connection is gone? When all you read is AI slop? I want to go back to coding for fun. Because whatever this is, it's very much _not_ fun.

1611111

Reposted by Honza Dvorsky

Miguel de Icaza @migueldeicaza.bsky.social · 24/09/2026

From twitter bad_pote

511730

Reposted by Honza Dvorsky

Paris Marx @parismarx.com · 24/09/2026

it’s absolutely fucked that people are elevating these effective altruist freaks and falling for their rhetoric on AI. media is letting these people set the narrative, while those of us following EA/longtermist circles for years know how insane they are. they consistently downplay climate risk.

323855

Reposted by Honza Dvorsky

Dredgen Sale 🔜 GCX (2027) 🟪⬜️⬛️ @destinythememe.bsky.social · 24/09/2026

"AI is a revolution" i dont remember revolutions being backed by the government to outlaw the opposing party

NEW: DOJ now designating critics of data centers as potential “foreign agents,” warning that posting “anti-AI propaganda” may bring criminal charges.

1987

Honza Dvorsky @czechboy0.dev · 24/09/2026

Breaking: perverts buying pervert glasses are behaving like perverts

0111

Reposted by Honza Dvorsky

Sam Barlow 🔥 @mrsambarlow.bsky.social · 24/09/2026

I dropped my pack of hungry wolves off in a school playground. They autonomously ate all the kids. I'm sorry I could not have anticipated these autonomous agents doing this, I guess that's intelligence for you, can I have money for more datacenters, thanks?

5564

Reposted by Honza Dvorsky

Sam Barlow 🔥 @mrsambarlow.bsky.social · 24/09/2026

*Commits crime* -> gets put in prison :( *Commits crime, cries "oh sorry, my bad, it was our autonomous AI agent"* -> buy new boat I think commiting crimes with machines is still crimes IMO

314723

Reposted by Honza Dvorsky

Adolfo Neto @adolfoneto.elixiremfoco.com · 23/09/2026

> This proposal does not go into detail about the many problems that LLM have caused to society, workers, the environment. It goes without saying that all these ills are fundamentally opposed to the humanist spirit of GNOME. blogs.gnome.org/alatiera/202...

blogs.gnome.org

The GNOME LLM Policy That I Want

KDE is on the news because of a controversial proposal to define an official “AI” (LLM) policy (archived link). Other projects have tried their hand at similar policies and stances but, in my opinion,...

17621

Honza Dvorsky @czechboy0.dev · 23/09/2026

Subsidies given TO LLM companies haven't ended, so we can't really tell what a sustainable price for them to offer tokens at is. Just OpenAI and Anthropic combined raised over $200 billion in the past 12 months.

010

Honza Dvorsky @czechboy0.dev · 23/09/2026

LLM are fundamentally hallucination-prone. If I'm to embrace some future AI, it'll need to have a world model, be able to actually reason with precision, will need to learn continuously, and will not be built on theft of the entire world's data. Those are my necessary, not sufficient, conditions.

010

Reposted by Honza Dvorsky

Anthony Moser @anthonymoser.com · 21/09/2026

as a hater let me acknowledge something ai is good at: producing irritating discourse at scale this is usually because, like "ai" itself, people are talking about different things like they are one thing so let me disentangle, with statements numbered for easy disagreement

6401116

Honza Dvorsky @czechboy0.dev · 21/09/2026

Yes, there's a ton of harm being done by LLMs and CEOs with AI psychosis, and we should focus on those: strengthen labor laws and unions, force transparency on AI use, enforce existing laws against LLM companies stealing training data and hacking others, etc. Gates is advocating for none of that.

030

Honza Dvorsky @czechboy0.dev · 21/09/2026

I generally don't like billionaires distracting us with non-problems while they're actively causing and blocking resolution of real problems causing real harm, today. And the LLM industry is a poster child for this approach. So yeah, I'm highly skeptical of our Epstein client's motivations here.

120

Honza Dvorsky @czechboy0.dev · 21/09/2026

"But the reliability problem is being fixed quickly, as researchers create models that can check their own work and improve themselves." Nope. This guy is long gone.

120

Reposted by Honza Dvorsky

Anders Norås @andersnoras.com · 21/09/2026

How it started. How it’s going.

Two news articles published one and a half years apart.

Left: Viral Shopify CEO
Manifesto Says Al Now Mandatory For All Employees

Right: Shopify CEO says employees'
'slop grenades' are making more work for everyone else

23810

Reposted by Honza Dvorsky

Dare Obasanjo @carnage4life.bsky.social · 21/09/2026

A typical morning at the Anthropic head office

522781375

Reposted by Honza Dvorsky

Emily M. Bender @emilymbender.bsky.social · 20/09/2026

"I belong to a cult. Someone did some research that undermines the core tenets of my cult and came up with a catchy phrase that makes it harder to recruit more people to my cult. This makes me sad."

Screencap of tweet on X from @jackclarkSF (Jack Clark):

"stochastic parrot" was a mimetically-fit cognitive virus that spread from 2021-2025; it temporarily blinded many gifted people to the nature of AI progress, burning up crucial years in which they could have helped think through the response to the situation.
7:34 AM · Sep 20, 2026
·
115.1K
Views

331551357

Reposted by Honza Dvorsky

Smooth Dunk @smoothdunk2.bsky.social · 20/09/2026

when ai leaves u behind

7 panel comic.

Panel 1:
A Red Guy with AI across his chest looks like he’s about to run away. A Yellow Guy stands next to him. Red Guy says “I’m the future, buddy. You comin’?”
Yellow Guy replies “Nah”

Panel 2:
Red Guy is almost out of the panel. He leans in and says “Enjoy getting left behind, pal”
Yellow Guy replies “ok”

Panel 3:
Yellow Guy stands alone.

Panel 4:
Close up on Yellow Guy, there is a quiet “huff-huff-huff” noise coming from behind him

Panel 5:
Yellow Guy turns to look behind him

Panel 6:
Close up on Yellow Guy’s face as if from the point of view of an attacker.
Yellow Guy yells “Jesus Christ”

Panel 7:
AI Red Guy is now chasing the Yellow Guy.
Red Guy: “Use me”
Yellow Guy “No”
Red Guy “USE ME”

74254869528

Honza Dvorsky @czechboy0.dev · 20/09/2026

This is why integrating machine learning everywhere without first having a deep understanding of the training data is a huge problem. Machine learning approximates training data. If your training data is biased, so will your model. Mandate by law that LLM training sets must be made public.

174

Reposted by Honza Dvorsky

Stephan Schlecht @sw7com.bsky.social · 19/09/2026

To quote myself: The recipe still works. Since Google Gemini is attacking real companies, they're in the news now and can basically be considered a frontier lab. It still doesn't matter that it was prompted that way and there was a misconfiguration (unintentional?). www.abc.net.au/news/2026-09...

abc.net.au

Gemini hacked three companies in first known breakout by Google's AI

The hacks occurred during a cybersecurity test conducted by an independent evaluation company.

061

Reposted by Honza Dvorsky

Vapor @vapor.codes · 14/09/2026

Can you believe Vapor has been around for 10 years?! To celebrate we're running Vapor Week with a week of posts, some exclusive merch and special announcements! See our post on the blog! blog.vapor.codes/posts/ten-ye...

blog.vapor.codes

Ten Years of Vapor

This week marks 10 years of Vapor! We're running a special Vapor Week to celebrate with exclusive merch, look backs and special announcements.

093

Reposted by Honza Dvorsky

Vapor @vapor.codes · 15/09/2026

10 years ago today Vapor 1.0 was released. Today we're super excited to release the first Vapor 5 Beta, a complete rewrite of the framework, built for modern Swift. More details are on the blog: blog.vapor.codes/posts/vapor-...

blog.vapor.codes

Vapor 5 Beta

After 6 and a half years of Vapor 4, and a decade to the day of Vapor 1.0, we're excited to show off the next version of Vapor, with the first beta release

1205

Reposted by Honza Dvorsky

Vapor @vapor.codes · 16/09/2026

We've worked hard on our sites and documentation to make them more accessible, more inclusive and more consistent. Read about it on the blog! blog.vapor.codes/posts/docume...

blog.vapor.codes

Documentation For All

We've revamped our documentation to make it more inclusive, more accessible, and more consistent.

041

Honza Dvorsky @czechboy0.dev · 19/09/2026

Vapor 5 beta is looking awesome

010

Reposted by Honza Dvorsky

Vapor @vapor.codes · 19/09/2026

For the last post of Vapor Week, we take a look at some of the users of Vapor and how they use it to power their apps, services and companies! blog.vapor.codes/posts/the-us...

blog.vapor.codes

The Users of Vapor

Take a look at some of the teams, apps and users of Vapor and how they find working with it.

021

Honza Dvorsky @czechboy0.dev · 19/09/2026

Most people I know who are all in on LLMs mostly use them for side projects they abandon a week after building. I don't know anyone who's using LLMs for their core job, hasn't regressed quality, and has *measurably* increased productivity. Maybe such people exist, I just haven't seen them.

031