Trying to make things that vaguely resemble games (and the tech to make them work). Recently released corpo/ghost as @InterleavedSys
Latvia
Joined August 2011
- Tweets8K
- Following89
- Followers175
- Likes25.5K
the subsequent posts are a thread on what the actual noise algorithms look like
makes it really clear just how far off the original (quoted) "AI" slop post was
people who don't know the subject matter shouldn't be allowed to ask chatbots
Arvīds Kokins retweeted
Freezers should probably not be connected to the internet.
I'm afraid we'll soon conclude that nothing should be connected to the internet.
Cyberattack on French supermarkers. Someone remotely took over refrigeration controllers and made them to defrost. Target temperature set to 50°C. Staff opening freezers in the morning were hit by a wave of hot air. Millions of euros lost. France spent years debating food sovereignty to depend on a firmware update? bfmtv.com/economie/consommat…
another data point to add to the "US is russia 2 now" folder
this one doesn't even need any deeper analysis
Trump administration refuses international election monitors for first time in over 2 decades.
The decision not to invite observers from the OSCE puts the U.S. alongside Russia and Belarus, while the Justice Department prepares to deploy its own poll monitors.
ms.now/news/trump-midterms-o…
and in the meantime they're also meddling in foreign elections - another thing they have in common with russia, though also not uncharacteristic for their own history
theguardian.com/us-news/2026…
Arvīds Kokins retweeted
Game studios are already signaling having to shut down their games due to the EU #KidsAct!
Sign #StopKillingInternet to keep our private servers safe from being illegal, and to keep age verification away from our games!: eci.ec.europa.eu/066/public/
If this proposal passes as written, it is the end of SuperTuxKart in the EU — and possibly the end of the project.
We need to be serious for a moment.
We cannot "block the EU market." We cannot prevent private servers. The code is GPL. Anyone can run one.
That is not a loophole — it is the entire point of open source.
We could not detect EU players on VPNs without tracking our players. We will not track our players. Not for Brussels, not for anyone.
A law written for tech giants is about to outlaw free and open source software made with love.
anthropic is led by illiterate nutjobs
books (especially ones meant for younger audiences) will often spell out the mental state of the characters
so will the people recounting their own experiences
that's the source of those "emotional vectors"
it's not magic
a rationally minded person would take their full, uncompressed dataset and look for structural similarities - words in the same order, very close together
each similar location should then be inspected to see what came before it (i.e. what phrases could trigger it)
Arvīds Kokins retweeted
Jesus fucking Christ
If you live in the EU and love videogames, please sign the petition here
eci.ec.europa.eu/066/public/
It takes like 2 minutes and it might help not live in an even worse shithole that we're already in
We just can't catch a break, can we?
The EU Kids Act is nightmare fuel, if you look beneath the honorable surface claim of "protecting kids online". The breakdown: youtube.com/watch?v=mQRcK9nY…
#StopKillingGames #StopKillingTheInternet
Arvīds Kokins retweeted
Connecticut’s latest privacy law, which bans companies’ sales of individuals’ precise location and the use of shoppers’ data to personalize prices, kicks in Thursday as industry attempts to block it persist. news.bloomberglaw.com/privac…
Arvīds Kokins retweeted
You've heard of AI-hallucinated cases and quotes.
Now try hallucinated witnesses and testimony from them!
From New Mexico, here's an order where a lawyer admitted ChatGPT wrote the briefs; made up cases, people, and testimony; and he didn't check any of it.
Arvīds Kokins retweeted
Ten days y'all then we get the unsealed order but this is HUGE. It's not generative AI specifically but it's about copyright and training and the affirmative defense based on fair use and had 6+ amicus briefs so IT WAS IMPORTANT. This was the first paragraph of Thomson Reuters brief. 1/2
Bam! For those who have tried to falsely suggest courts have determined AI training is covered by fair use, not only do we have Judge Chhabria disagreeing in a motion to dismiss, we now have Thomson Reuters v. Ross Intelligence upheld by the third circuit. reuters.com/business/media-t…
I see some big problems here:
- 15 minutes for both tests? those are tiny tests (quizzes actually)
- testing + lesson in one 75 minute period?
- AI lesson at home? so nobody was watching them (e.g. to see if they don't cheat)?
- can't find the text of the tests and if pre=post
Researchers proved AI has deleted every reason universities exist.
Harvard University ran a controlled experiment pitting a custom AI against their own top-tier classrooms.
And the results are going to collapse the higher education bubble.
They took 194 undergraduates and split them up.
One group learned physics in one of Harvard’s best hands-on, active-learning physical classrooms. Group work. Instructor support. The premium university experience.
The other group went home and learned the exact same material with an AI tutor.
The AI didn't just win. It embarrassed the institution.
Students using the AI learned more than twice as much as the students in the elite Harvard classroom.
They scored 30% higher on the final assessment.
And they did it in less time.
Let that sink in.
A piece of software sitting on a laptop outperformed a world-class faculty in one of the most elite learning environments on Earth.
Universities have always justified their exorbitant tuition with two things: access to elite knowledge and the physical classroom experience.
This study just proved both of those moats are gone.
When software can teach you complex physics twice as well as a $60,000-a-year institution, the math of higher education breaks permanently.
The AI didn't just give the students answers. It used strict pedagogical guardrails. It guided. It questioned. It forced the students to do the cognitive work.
It offered perfect, one-to-one tutoring, personalized to the exact moment a student misunderstood a concept.
That level of attention is mathematically impossible to scale in a physical lecture hall.
For a thousand years, the university was the only place to get a premium education.
Now, it’s the bottleneck.
If AI can double your learning speed for a fraction of the cost, what exactly are students taking on decades of debt to pay for?
the conclusion is that there's no reason to use this as some proof that "AI" outperforms in-class learning
letting the study participants do the post-test (quiz) at home is the biggest red flag
also, for some reason they did a whole news segment for it: nbcnews.com/now/video/harvar…
this is just a common sense request that many companies made internally back in 2023 already
I'm surprised at the defensive people in replies either ignoring the second paragraph of the first image, or saying that some "do not train" toggle is enough to protect one's data
Arvīds Kokins retweeted
Anthropic's discovery of array-associated reverse transcriptases?
Turns out it was already done by a Danish scientist. He used Claude for administrative tasks, but not in an agentic manner for the discovery. He already filed a patent and gave talks on it
nytimes.com/2026/09/27/scien…
consequently, the idea that gen"AI" can "earn" its value? if the value is in fact negative? technically it is earning its value now by losing money
the vast funding spent on the GPUs and the construction of new "data centers" is all the value gen"AI" will ever create
more broadly speaking, anything you can convince someone to pay you for expands the economy
and anything causing the opposite shrinks it
economy = how many different ways of being useful to each other we have figured out
well, that, plus any extortive payments (rents)