Hans-Christoph Steiner @eighthave@librem.one
Joined Aug 2019
oh funny someone published an academic research paper on this, the abstract sounds interesting: https://www.nber.org/papers/w34910
anyone know anything about it?
Could there be an "uncanny valley" of #AI #LLM #bullshit, where the error rates are low enough that humans get lazy and stop checking, which then leads to higher error rates than when humans did it without AI. Unfortunately, the current #BigTech business models prioritize #adoption and #engagement, so that will bias them towards ignoring this affect.
Hans-Christoph Steiner
boosted
This is absolutely nuts: hackers are hijacking high-profile Instagram accounts by simply asking Meta's AI chatbot to change the email on the account. Meta's AI does it, hacker gets password reset code, they're in. A staggering security issue
Hans-Christoph Steiner
boosted
RE: https://social.wildeboer.net/@jwildeboer/116679473218827225
I'm so proud of our government institutions that are moving to Mastodon 🥹
There are now several Dutch ministeries, provincies, agencies and municipalities, but also officials such as state secretaries and ministeries on social.overheid.nl 🥳
A big thank you to @manfredzielinski for making this happen.
And what I find quite funny is that maybe a third of the phone call would be taken up with the standard Austrian pleasantries one must use, including a proper greeting and minimum two forms of goodbye 😉
Someone should do a quantitative #productivity analysis of #Debian's #workflow vs #Microsoft's for #Windows, or #Google's for #Android which both surely include #LLM #AI tricks that this point. I'd bet both those companies have done that internally. Any chance of leaks? 😉 So much of how Debian works is in public, so anyone could do some kind of analysis.
I've been an early adopter of #Internet things for decades, but at the same time, I work to be able to realistically assess the value as compared to other methods. Many things in #Austria still happen in person and via the #telephone, so its a good testing ground. I'm frequently surprised at how #efficient in-person/telephone can be when compared to #online, like government offices have employees that handle making appointments by phone, they are often really good at it, so its surprisingly fast
The #internet is full of fun locking #garden #hack ideas for all sorts of things but almost always, they use #toxic materials that you really don't want leaching into your #soil. Even common #plastics like water bottles or #polypropylene turn quite toxic when exposed to #sunlight for a while. For example, https://rosal.web.uah.es/pub/Aquatic%20toxicity%20of%20commercial%20polypropylene%20and%20associated%20chemicals.pdf
Hans-Christoph Steiner
boosted
No matter the geography, a common reality emerges: those least responsible for the crisis are on the frontlines of confronting it. In doing so, they are not only protecting their territories, but safeguarding the ecosystems that sustain life on Earth, often at great personal risk.
Take action and stand with Indigenous Peoples to protect their territories.
Link in bio.
Hans-Christoph Steiner
boosted
My personal conclusion can however not end up with anything else than that the big hype around this model so far was primarily marketing. I see no evidence that this setup finds issues to any particular higher or more advanced degree than the other tools have done before Mythos. Maybe this model is a little bit better, but even if it is, it is not better to a degree that seems to make a significant dent in code analyzing.
Joined Aug 2019
