Chin up! This was the first community I added when I join Lemmy…and I appreciate you moderating it.

I do think it’s relevant / I think we’re arguing the same point from different ends. I’m simply underlining the idea that this whole shooting match isn’t AGI and that we should push back on that framing, before we solve the wrong problem in the wrong way.

It’s still dangerous, but semantic games of gotcha are not what needs to happen.

The app.

Fair correction. I should have been more specific and not dumbed it down for lurkers / lay audience.

My broader point though is neither that (nor the medicare “hack”) is proof that skynet is knocking on the front door.

OTOH, the agents are displaying certain emergent behaviors that are worth mulling over - like the obfuscated back channel communication. Not “sapient” (and I use that word deliberately) but … curious / problematic, from a control pane level.

Well, in the sense that neither was evidence of runaway AI becoming sapient and deciding to act maliciously, which is often how the “AI hacked X” framing gets presented.

Yes, the Hugging Face incident was more technical / concerning. (For those not in the loop: OpenAI agents in a cyber-security eval sought answers outside the eval environment, found useful material on Hugging Face, discovered HF credentials, and then tried to exploit them. Basically, the agents were trying to cheat on an exam.)

My point was about the narrative around how the media presents “AI hacking”, not that the two incidents were technically equivalent.

Hyperbole, but not by much.

I’d screenshot the app for you but apparently that’s blocked…and it’s (once again) not accepting passwords / down at the moment.

Right. So this may have been less “AI defeated Medicare security” and more “the agent found a weakness that a human could also have found and used”. For all we know, someone left a plaintext password somewhere accessible and the agent found it with a simple search.

Until the technical details are public, we don’t actually know what this supposed “hack” involved. Though “Medicare defeated be CTRL+F” would by chef’s kiss after RoboDebt.

A cynic might also wonder whether this becomes another wedge for more identity and age checks online. I already got pinged earlier this week with the ridiculous “papers, please” when trying to access YouTube.

I’d rather know what actually happened before we decide that the lesson is “everyone needs more authentication” or “Skynet is here”. The whole thing has a familiar stank to it, much like the supposed Hugging Face “hack” from the other month.

I’ve seen a few of these articles in passing and I am still unclear - what was the actual “hack”? What did the AI agent actually do, in the technical sense? There is a world of difference between it doing a stupid brute-force attack and it discovering and exploiting an access-control flaw that the site itself exposed.

Neither is good…but if Medicare basically left the front door open, it wouldn’t take more than a bored 13 yr old to do the same. Let’s pump the brakes on Skynet.

BTW, can we talk about why we still need 19 factor authentication to log into My.gov? That’s always a hoot and the experience is always smooth. Gee, I wonder whether the same committee that designed the myGov login experience might also have made some questionable security decisions elsewhere?

Interesting corollary for you. I had an old google account with an anon email address (skiff.com, I think it was).

When I signed up, they accepted it, but in the intervening time, decided no-mas.

For the longest time, I couldn’t post comments, upvote etc on Youtube (no great loss)…until I signed up with a new account using a @proton address.

Wow. MAC address targeting maybe?

…

What did you DO to piss off dear old Google that badly? :P

You’ve been IP range blocked by sounds of it. I did a little reading over the weekend and found that non logged in users (eg: someone using yt-dlp) can and do get rate limited / temp banned.

https://reddit.com/r/youtubedl/comments/1ryxhdk/youtube_login_requirement_after_using_ytdlp/

The duration seems to be variable - between hours and days / weeks.

Ok but the claims made by the article are factually wrong, so… 🤷

None of those - searxng, deGoog, 4get - are private search engines. The are metacrawlers / indexers for things like Google, brave, DuckDuckGo, etc etc.

The difference is not subtle.

They anonymise your request to the underlying index, but the index they’re querying is still Google/Brave/DDG etc…so you can still be fingerprinted over time.

If you want a private search engine, the only one that comes to mind in YaCy (which is peer2peer).

If you don’t want to self host or p2p, your options are Mojeek (own crawler, index, no tracking), Marginalia (ditto but very niche) and maybe Kagi (which is sort of hybrid).

sigh.

Here we go again …

“It feels like a cheat code to me. It feels like I’m not going to be as smart as the people who are currently in the field that I want to be in — because of AI.”

1. Em dash in a complaint about em-dash-loving clankers. Hmm. Is it clankers all the way down, ABC? Should we just talk to ChatGPT directly and bypass all pretense?

2. Nothing requires her to outsource her thinking to AI. She can develop critical-thinking skills and use AI for grunt work instead. So, that is either a failure on her behalf or of those teaching her. In fact, she conceded that point herself later in the piece (“…AI is a tool, but it’s being used in the wrong way.”)

But if you use ChatGPT to do the thinking, then you practise those skills less! - yeah, no shit? That’s a problem with how the tool is used, not proof that the tool makes people stupid.

More to the point, there’s a growing body of research showing that social media (which predates gen AI…and whose use is endemic in her age cohort) has resulted in poorer memory, reading, vocab and cognition.

https://jamanetwork.com/journals/jama/fullarticle/2839941

https://pubmed.ncbi.nlm.nih.gov/41231585/

Additionally, the whole “digital native” idea has been on shaky ground for years.

https://www.sciencedirect.com/science/article/abs/pii/S0742051X16306692

https://www.sciencedirect.com/science/article/abs/pii/S0360131518301738

Being extremely familiar with phones, apps and interfaces does not automatically produce broad digital literacy or general-purpose computer skills. Meaning, while older generations were coding in BASIC and disassembling RC cars for fun, the large bulk of her cohort was distracted by Insta and Snapchat.

Quelle surprise they didn’t develop background skills going into technical fields.

TL;DR researchers were documenting the gap between familiarity with consumer technology and actual digital competence years before ChatGPT existed…and that was BEFORE factoring in the brain-rot of social media.

Blaming AI for a pre-existing digital-literacy gap (and cognitive decline) gets the chronology backwards. Dare I say, it shows the absence of those very same critical thinking skills she rightly claims not to have developed.

PS: The enviro claims are overstated too.

https://blog.andymasley.com/p/a-cheat-sheet-for-conversations-about

Pearl clutching journalism from the ABC - say it ain’t so!

Postscript:

Like many, I prefer having an offline cache of YouTube videos, because things disappear, channels get taken down etc.

Back-of-the-envelope maths suggests something in the order of 1TB could comfortably hold ~1,500 720p videos averaging around 20 minutes. That’s a pretty decent little seed library, without fossilising the whole thing into YouTube2ZIM/Kiwix.

A lot of self-hosters use TubeArchivist, or dump downloaded videos into Plex/Jellyfin, but I’ve never found either approach particularly satisfying.

So, I’m going to see whether I can make my SmartTube fork also consume actual offline-downloaded YouTube videos as first-class content.

That way, you could simply feed it downloads produced by something like ytdl-sub.

If there’s interest in that, please upvote / say so.

That sounds like PeerTube with extra steps :) I’m down for it.

I’ll probably end up using Pinchflat over Tube Archivist myself (a bit lighter to keep running as always on service on my Raspberry pi) but p2p seeding should work just the same.

Yep, that’s a good point. I actually have a battle plan in place (if it comes to that) using a mixture of 540-720p videos (to reduce download sizes) and then AI upscaling as needed using an NVIDIA shield or on site compute.

They really, really don’t like it when you don’t log in, huh? 🤣

Fun game of cat and mouse aside, if it all ever goes to shit…I’ll miss niche YT channels but there are other alternatives.

Odd. Newpipe and Smarttube (the one I just cooked) all seem to work here. I wonder if they’re doing IP range blocking now in certain geos.

When you say “prove you’re not a bot” - do you mean they force you to sign in or is it just a CAPTCHA?

Hmm. Can anyone else confirm similar issue to @muhyb@programming.dev

What? Really? Embedded videos are NOT public anymore?

That’s…not great. For multiple reasons, but not least of which m.youtube.com likely breaks.

Well, I expect a sudden upswing in Tube Archivist will be forthcoming