Sabitlenmiş Tweet
Patrick's Law
"An organism can only create artifacts
which are less intelligent than itself."
English
CryptoPatrick
330 posts
CryptoPatrick
@cryptopatrick
Mathematics Undergrad @ChalmersUniv 🇸🇪🦀 Digital Economics | Ai | Seamless Global Systems
Katılım Mayıs 2015
457 Takip Edilen207 Takipçiler
@L42ARO @Saboo_Shubham_ I think that people who make serious money
tend to avoid positioning themselves
on either side of right or wrong.
English
“OpenClaw is the iPhone of tokens” — Nvidia CEO on Lex Podcast
English
The iOS app (v1.3.51) and X version run on separate inference paths: mobile often uses a lighter/quantized snapshot for speed/battery, while web hits the full latest backend. That decimal ordering glitch looks like a classic parsing slip (lexical vs numeric), not human-added guardrails. Deployments aren't perfectly synced even in 2026—version lag + prompt tweaks explain the diff far better than patching. SOTA models still trip on edge math occasionally.
English
Date: March 25, 2026
What: Grok app, version 1.3.51 on iOS clearly hallucinating when ordering reals.
@grok is the fact that the iOS version of Grok hallucinated, while the twitter version did not, a possible sign of patching in the form of deterministic guardrails added by non-AI agents? Theorize on why the two answers differ on what is assumed to be SOTA LLMs for each platform, at todays date.
Dragonfang@_Dragonfang
@cryptopatrick @mgubrud @MaMoMVPY Please name a single frontier model which claims 1.9 < 1.11. We're talking about the present, not the past.
English
@L42ARO @Saboo_Shubham_ A new interface that everyone will want to use (multi-touch), and a new platform which everyone will want to build for (mobile-first).
English
@F33dc0de @jasonbosco You can take pride in being a lumber jack with an axe, but that's not how logging is done today.
English
@jasonbosco @cryptopatrick Taking pride in not looking at the code is misplaced pride and it will come back to bite them. I still do manual code reviews even with LLM coding most of my stuff
English
I see a new form of tech debt coming for dev teams -
Comprehension debt.
As more and more code is generated by LLMs, if teams don’t take the time to understand deeply what the generated code is doing, as well as code they write by hand…
It’s only a matter of time before the code base starts looking unfamiliar to most of the team.
It then becomes harder to discern if new code that LLMs generate is adding more spaghetti or if there’s a better approach.
It’s a downward spiral from there - unrelated things break with every change despite existing tests passing, no one knows the full picture to be able to fix the root cause, not even an LLM, etc.
So as tempting as it is to move super fast with LLMs, there’s only so much comprehension debt you can rack up before your code base silently becomes a Rube Goldberg machine under your nose.
English
English
@_Dragonfang @mgubrud @MaMoMVPY @grok So, it kind of looks patched!
🤷♂️👇
"patching cases of hallucination as they are discovered feels more like curve fitting, than what most people expect when they hear AGI."
English
A year ago Jensen was a completely reasonable sounding CEO. Now he and the rest of the AI CEOs are just making up obvious fantasies about the state of AI.
We have not achieved AGI. He knows that we are nowhere near that.
Yuchen Jin@Yuchenj_UW
Jensen: “I think we’ve achieved AGI.” “It’s not out of the question that an OpenClaw could create a web service or some interesting little app that, all of a sudden, a few billion people use for 50 cents.”
English
@adxtyahq @AdnanShams_ @amritwt @archiexzzz How do people know if they are ready to apply to their 1st internship?
I'm building everday, but an internship sounds kind of scarry to be honest - everyone seems so cracked. 🫤
English
English
I might sound unrealistic or retarded but is it really possible to get a job/internship in 6 months if someone starts from scratch today???
@amritwt @adxtyahq @archiexzzz
English
@Anthony_Pugh2 Making 100 contested shots AFTER running a mile AND doing 100 pushups > MAKING 500 shots a day
English
MAKING 500 shots a day > SHOOTING 1,000 shots a day
Just my opinion…
English
@kimmonismus What's your opinion on LLMs and hallucinations - with reference to AGI?
x.com/cryptopatrick/…
CryptoPatrick@cryptopatrick
When you say "we have achieved AGI." are you specifically referring to SOTA LLMs? If so, how does the tendency of these LLMs to hallucinate on some tasks, tasks which grown humans can typically solve with ease, affect your claim? Would you fly with an airline whose pilots claim with conviction that 1.9 < 1.11? As you well know, these hallucinations are a direct consequence of how LLMs (assuming that's what you're reffering to) function at a fundamental level - patching cases of hallucination as they are discovered feels more like curve fitting, than what most people expect when they hear AGI.
English
@thePartyPartyUS @jimstewartson Looks like you might have been right.
The inventor of the AGI term (of that's true) claims that AGI has been achieved! It will be interesting to see/read how he has confirmed AGI.
Ps. I appologize for calling you a dumbass - I shouldn't have done that.
Have a great week, sir!
Mark Gubrud 🇺🇸@mgubrud
@MaMoMVPY Well, Lars, I INVENTED THE TERM and I say we have achieved AGI. Current models perform at roughly high-human level in command of language and general knowledge, but work thousands of times faster than us. Still some major deficiencies remain but they're falling fast.
English
@cryptopatrick @jimstewartson look you're just bad faith time wasting at this point.
I appreciate you're a student. I appreciate the desire to formalize. I appreciate people building things.
Your law is nonsensical and I'm sure you'll find that yourself over time. God bless.
English
JUST IN: Jensen Huang is the central conspirator in the largest fraud in the history of mankind.
We have NOT “achieved AGI.” We will never, ever achieve AGI with LLMs—and he knows it.
Trillions are being LIT ON FIRE.
This is not technology. It’s an elite Ponzi cult.
Watcher.Guru@WatcherGuru
JUST IN: Nvidia $NVDA CEO Jensen Huang says "we've achieved AGI."
English
Interesting, @mgubrud claims that AGI has been achieved.
I'm anxious to see if it's true, and how it's been confirmed.
Has Patrick's Law been refuted?
⏳
Mark Gubrud 🇺🇸@mgubrud
@MaMoMVPY Well, Lars, I INVENTED THE TERM and I say we have achieved AGI. Current models perform at roughly high-human level in command of language and general knowledge, but work thousands of times faster than us. Still some major deficiencies remain but they're falling fast.
English
Patrick's Law
"An organism can only create artifacts
which are less intelligent than itself."
English
@BuildTestShipAI @OmriBuilds Thanks for the kind words. 🙌😀
The grammar projects are related to declarative AI and have little to do with what you describe.
English
@cryptopatrick @OmriBuilds I love the video game concept.
How do the grammar projects work? Currently working on a history app that involves quite alot of text and would love to improve the writing quality before launch.
English
@thePartyPartyUS @jimstewartson The definition of an artifact?
English
@cryptopatrick @jimstewartson i want you to provide the definition :)
English
When you say "we have achieved AGI." are you specifically referring to SOTA LLMs?
If so, how does the tendency of these LLMs to hallucinate on some tasks, tasks which grown humans can typically solve with ease, affect your claim?
Would you fly with an airline whose pilots claim with conviction that 1.9 < 1.11?
As you well know, these hallucinations are a direct consequence of how LLMs (assuming that's what you're reffering to) function at a fundamental level - patching cases of hallucination as they are discovered feels more like curve fitting, than what most people expect when they hear AGI.
English
@MaMoMVPY Well, Lars, I INVENTED THE TERM and I say we have achieved AGI. Current models perform at roughly high-human level in command of language and general knowledge, but work thousands of times faster than us. Still some major deficiencies remain but they're falling fast.
English
@thePartyPartyUS @jimstewartson The discussion:
AGI and the creation of artifacts.
Enter you:
"What about babies dumb dumb."
What do you want me say
when you open with something
which has nothing to do
with the discussion?
English
@cryptopatrick @jimstewartson i said there is no definition of artifact you can give, and you agree, because you shifted back to talking about babies. There are other examples I can give that would violate any definition of artifact you would provide. you know this, or you would offer one. 'Law' lol
English
@thePartyPartyUS @jimstewartson I don't have time to convince you, that people world wide do not include human babies in the artifact concept.
Maybe try pointing at someone's baby and say "Your artifact is very cute!", see what kind of reaction you get. 😂
English
@cryptopatrick @jimstewartson There is no definition of artifact you could give that would save the argument, fool
English