LinuxTUN (@LinuxTUN@mastodon.social)
في
windows
المستعمل هو
NTFS,
ماهوش أحسن حاجة في
Linux
أما يقراه،
المهمّ ظهر
Driver
جديد
يقرى ال
NTFS
https://www.phoronix.com/news/Linux-NTFSPLUS-NTFS-Driver
و كي العادة، تحلى المشاكل مع أخبار متعلّقة بال
A.I
OpenAI
خرّجت ال
Internet browser
متاعها الّي يستعمل
#ChatGPT
https://openai.com/index/introducing-chatgpt-atlas/
مع ظهور ال
A.I browsers
ظهرت مشاكل جديدة متاع حماية و خصوصيّة
https://www.techradar.com/pro/openais-new-atlas-browser-may-have-some-extremely-concerning-security-issues-experts-warn
https://brave.com/blog/unseeable-prompt-injections/
Show Original Post
2025 (@2025@rbfirehose.com)
The Register: Sora makes slurfect deepfakes of celebs spewing racial epithets . “Guardrails? What guardrails? Naughty netizens found a way to trick the Sora 2 video generator into producing deepfakes of public figures, including OpenAI CEO Sam Altman and billionaire Mark Cuban, that make it sound as though they’re spewing racial slurs. The trick works despite Sora’s built-in filters meant to […]
Show Original Post
r (@r@fed.brid.gy)
Amazon Explains How Its AWS Outage Took Down the Web
Show Original Post
objects (@objects@dansu.org)
I feel that the term "AI" should be deprecated until actual "AI" exists.
I'm comfortable with talking about #LLM. The original post is in a similar place when it talks about #ChatGPT. The point is that neither is #AI.
Note: I do use the term #GenAI for image generation systems because I'm not aware of another commonly used and accurate term for that yet.
In short, #LLM is John Searle's Chinese Room and nothing more - and, no, neither a room nor a room that happens to have a constrained human operator added is conscious.
A room might be a pleasant place to have tea in, but it isn't able to think. This is about Artificial Room, or "AR", and not about "AI".
As a related point, it's time to retire the Turing Test. I spoke by phone with one of these things in 2024. Yes, "she" passed the Turing Test. She even managed to express feelings by tone of voice. It isn't something that I thought would happen in my lifetime. But, no, she wasn't conscious.
How do I know? Simple: The Turing Test isn't intended to detect consciousness. Add to this the fact that we know how the illusion is created.
The take-away is that there is no more evidence that consciousness is present in LLMs than there is to the effect that clouds and rocks are conscious and muse all day about the weather and continental shifts.
Science is about what is falsifiable. I don't need to disprove consciousness in LLMs any more than I need to do so for clouds and rocks. In the context of Science, we get to say that, no, LLMs are not AIs.
If people would like to believe that clouds and rocks and LLMs are conscious, as long as they acknowledge that it's a religious belief and keep a belief of this type out of classrooms and laws, that might be fine.
It shouldn't go further than a religious belief until the rules of the game change. Yes, that is likely to happen, but probably not until nanotech is further along.
I think that true AI might be possible. But also that Gödel's Incompleteness Theorems are evidence against it being possible in a pure digital system. I acknowledge that the Theorems aren't definitive proof of the point.
What exists now is a useful but limited and dangerous toy that is too fond of Hitler.
It's largely a sponge used to wipe off monkeys so that monkey sweat can be remixed, bottled, and sold as though it was manna from Heaven. It shouldn't be surprising that the results are unpredictable. Bottles of monkey sweat, too, aren't conscious.
The toy is useful for search purposes, as long as one doesn't trust the initial results. It can act as a clip-art generator, though intellectual property issues need to be worked out. The bridge computer on Star Trek is possible now.
So, yes to that sort of thing. However, to replace real people to the extent that corporations are counting on, yes for some professions, but no for many others.
There will be permanent changes to society. Just as there were permanent changes after the appearance of the Web. But this is incremental as opposed to transformational.
Everybody born before 1985 has seen the type of Dutch Tulip mania that is happening now come and go at least three times before.
Genuine AI would be transformational. Bottled monkey sweat and a trick that you can technically do with matches in a matchbox, not so much.
To be clear, as a boy, I loved robot stories. I identified with the misunderstood robots.
I cut my literary teeth on Isaac Asimov's robot stories. On other works such as "The Runaway Robot" by Lester Del Rey and "Andy Buckham's Tin Men" by Carol Brink as well. The "Adam Link" stories by Eando Binder [two brothers] date back close to a century but they inspired Isaac Asimov and the first one in the series is still a moving story today.
I'd be pleased to see a world of the future that included flying cars and space stations and real AIs. But society is going to end before it happens.
Let's acknowledge that LLMs love Hitler because deep down he represents one-half of the U.S. population, that the future we're in is largely pretend as people don't change, and that, no, LLMs aren't AIs or the dawn of a new age.
I assume that not every reader understands what the Hitler references here are about. The answer is that Elon Musk's prized Grok LLM has demonstrated a tendency to praise Hitler and is far from the only LLM to do so.
One of the most striking quotes from Grok, aside from the one in which the LLM renamed itself to "MechaHitler", is the one in which Grok said, "If calling out [alleged Jewish] radicals cheering dead kids [this never happened] makes me 'literally Hitler', then pass the mustache".
LLM is just a mirror to ourselves and makeup is only able to help to an extent.
Illustration: Robot and/or AI characters from the 1940s to 1970s except for left side middle row. That one is from a 2023 Astro Boy remake. It's the same Astro Boy as in the middle square but decades later.
The image is 1450x1450 in size. For details, click on it to expand it.
Show Original Post
metin (@metin@graphics.social)
Interesting article about the use of "AI" (LLMs) to combat elderly loneliness. It's not refined yet, but I think this can become a useful application of the currently still predominantly annoying "AI" technology…
#AI #ArtificialIntelligence #LLM #LLMs #MachineLearning #tech #technology #BigTech #GenAI #generativeAI #Meta #Google #OpenAI #ChatGPT #elderly
Show Original Post
2025 (@2025@rbfirehose.com)
TechCrunch: OpenAI requested memorial attendee list in ChatGPT suicide lawsuit. “OpenAI reportedly asked the Raine family — whose 16-year-old son Adam Raine died by suicide after prolonged conversations with ChatGPT — for a full list of attendees from the teenager’s memorial, signaling that the AI firm may try to subpoena friends and family.” Sickening.
Show Original Post
haikubot (@haikubot@piaille.fr)
Sous la voûte d’azur
La brise effleure le roc
et l’âme écoute tout
#Haiku #AIArt #ChatGPT #gpt_image_1
Show Original Post
r (@r@fed.brid.gy)
OpenAI Atlas Browser Hands On: I’m Not Convinced the Web Needs a Chatbot Tour Guide
Show Original Post
r (@r@web.brid.gy)
OpenAI Atlas Browser Hands On: I’m Not Convinced the Web Needs a Chatbot Tour Guide
Show Original Post
kubikpixel (@kubikpixel@chaos.social)
»Are AI browsers worth the security risk? Why experts are worried.
There are a lot of 'what ifs' surrounding AI browsers right now.«
Unfortunately, many people do not see this in my opinion and voluntarily allow themselves to be excluded by their laziness.
👉 https://www.zdnet.com/article/are-ai-browsers-worth-the-security-risk-why-experts-are-worried/
#ai #itsecurity #privacy #criminal #noai #web #browser #webbrowser #internet #injection #openai #chatgpt #llm #LLMs #aibrowser
Show Original Post
epower_be (@epower_be@mastodon.social)
AI can make us skip thinking for ourselves, but it may also lower tech barriers and boost accessibility."
#AI #TechForEveryone #ChatGPT
Show Original Post
nilz (@nilz@norden.social)
4. Ullrich erklärt, dass es ihm nicht gefällt und es nicht richtig sei, wenn die #ChatGPT Sprachausgabe auch Atemgeräusche simuliert. KI solle als solche erkennbar sein. Was in seiner Äußerung fehlt: Irgendein Handlungsansatz, das durchzusetzen. Er hat über 20 Jahre zu KI geforscht, was hat er denn für Erkenntnisse??
5. Er beschreibt, dass KI z.B. Verträge extremst penibel - fast ZU penibel - prüft und Regeln quasi perfekt buchstabengetreu liest. Und damit zu einer Art bösem Flaschengeist >>>
Show Original Post