ptidhomme

Thanks for these insights. They made me think of AI as what a mechanic is for your car. He knows more and is more capable than you, and he can save you a lot of time vs if you had to do the dirty work yourself.

However, the more you know yourself about car maintenance, the more you will trust his diagnosis, solutions and the fairness of the money he asks for it. More generally, the more you will trust him to act in your interest instead of only his (or someone else's).

The edge in the AI future will be ownership.

The AI overlords will eventually consider the majority of human lives to be worthless. Let's face it, one day we will need to fight for mere survival.

Did you feed OpenAI models with your insights though ?

AI will eventually get better than you at all you do, because it trains on your inputs.

Secret knowledge/data will be tomorrow's gold.

"Geiger"

Prepping for AI apocalypse I see

Some Jews see AI as a messianic entity, so it fits the bill.

Interestinly, it does include the string "myrtus", which was likely it's real internal name.

https://medium.com/@arpit.writes/stuxnet-and-the-birth-of-di...

Well, it's actually an essay against the thinness of AI slop.

> Erasing the line between the thick and the thin has left us defenseless against slop at the exact moment of its onslaught. Everyone can sense there’s something amiss with the prose that comes out of the machines, but we lack the language to talk about it, and so we’ve converged on the idea that slop simply means using too many em dashes, bullet points, and line breaks.

Man, thanks for this precious tip

It's all parallel construction.

https://en.wikipedia.org/wiki/Parallel_construction

Is porn birth control ? Yes it is. But why is porn free and ubiquitous ?

Surely there's an enormous amount of money behind it, but where's the ROI ?

Someone must be funding the NGOs organizing it all.

Yeah, I also use it because it is fairly low maintenance. There's the sysupgrade every 6-month, but it goes smoothly every time.

I used to have the opposite rule in my signal processing field : the more Chinese names, the less innovation was there.

They seemed like they had to be churning out papers and any little adaptation to existing research triggered a new publication.

But it may have changed now.

I liked this reply in the thread :

There's another issue surrounding developer skill atrophy or stunting that I find \ particularly concerning on an existential level.

If we allow people to use LLMs to write code for a given project/platform, experience \ in that platform will potentially atrophy or under develop as contributors \ increasingly rely on out sourcing their applicable skills and decisions to "AI".

Even if you believe out sourcing the minutia of coding is a net positive, the \ "enshitification" principal in general should give you pause; as soon as the net \ developer skill for a project has degraded to a point of reliance, even somewhat, I \ think we can be confident those AI tools will NOT get less expensive.

I'd rather be independently less productive, than dependent on some MegaCorp(TM)'s \ good will to rent us back access to our brains at a fair price.

- achaean

https://marc.info/?l=openbsd-tech&m=177430829313972&w=2

I'm not sure why this would force him to retire as an OpenBSD developer ? Did this research bring him some backlash from other developers ? If not, maybe he can advocate for a better acknowledgment of the issue from the inside ? Just wondering.

Plus you can install them in a separate work profile (use app "shelter"), which isolates them more.

Personally I need them for my banking app (which is a shame but here we are), but I disallow network access for Google Play Services + disable Google Play app. The bank app works nice with that.

The theory still emanated from actual observations, didn't it ?

Now of course, the only input LLMs have is human text (for text only LLMs anyway). So their model is entirely dependent on how we see the world. I wouldn't restrict LLMs to description of human understanding. They can articulate concepts in a rather sensible way, that wouldn't exist as is in the training corpus. Which exactly means that they have a model, however limited or imperfect.

What would a human say about what he/she is or how he/she works ? Even today, there's so much we don't know about biological life. Same applies here I guess, the LLM happens to be there, nothing else to explain if you ask it.

Tbf banks are truly driving the coercion. My banks used to have functional apps on bare GrapheneOS, but nowadays they won't launch without Play Services (to wich I can deny network access, pretty neat of GrapheneOS). Plus my main bank now requires the app even for web access. Nightmarish.

"How concepts relate" is called a model. That it uses language to be interacted with is irrelevant to the fact that it's a model of of a worldly concept.

What of multi modal models according to you ? Are they "models of eyesight", "models of sound", or pixels or wavelengths... C'mon.

When you ask an LLM a question about cars, it needs an inner representation of what a car is (how imperfect it may be) to answer your question. A model of "language" as you want to define it would output a grammatically correct wall of text that goes nowhere.

A model of language is a model of the world, else it being pure gibberish.

Those billion parameters, they are a model of the world. Autocomplete is such a shortsighted understanding of LLMs.

Aren't botnet targeting cheap and unsecured consumer devices specifically in the residential IP ranges ?

GrapheneOS, OpenBSD, Wireguard

Filesystem is one of the rare areas (albeit crucial) where OpenBSD is flaky, tbf.

> "how do I get <thing> working"

OpenBSD is very different from FreeBSD in this regard. OpenBSD mostly works out of the box.

Man, I remember when you told this story some years ago, and I still very much like it ! (I have an hnrss.org feed with all comments mentioning OpenBSD, so, I was bound to catch it).