Zoe Kleinman’s Post

Cc Zsolt Virag-Tulassay

Magic , like AI, is very entertaining and even has shows at Las Vegas…… But Magic has no place in NHS surgery - as you cannot saw lady in half on the stage - you can fool an audience….. but not patient. Smoke & mirrors cannot compete with scalpals and reality bites. AI is entertainment IT - it is not mainstream IT (like databases, operating systems and applications) Here endth todays lesson, there is Magic circle chasing me…

Bang on the money. I don't think anyone is expecting Hal to lock us out of Discovery One anytime soon, but the capabilities for plagiarism bypassing consent is a very real threat.

Zoe Kleinman Very interesting long-form article, much better than a 5-min byte-sized slot on BBC Today program…. Key quote for me: “The independent report also states that there is currently no reliable way of understanding exactly why AI tools generate the output that they do – even among developers”….. oh dear Without understanding why AI does x, we cannot test x and therefore cannot support x and should not release x. This has become a ‘have blind faith’ in AI… no way its computer stuff and computers are not as reliable as humans in a panic or crisis! We are looking at Magicians for example…. Its great entertainment….and sells a few magic kits and can do show at las vegas - but NHS cannot cut lady in half….. its a trick and not a useful business venture. Sadly #postofficescandal shown £1 billion HorizonIT could not be trusted to add up 1 million transactions per day and get right answers - its generated discrepancies for 20 years which PostOffice belived were stolen by 700+ innocent subpostmasters - who were convicted by computer - and Terminated! Dont trust computers more than you can throw them…. Awooga Trust scale: Mobile phone will throw 10 meters Laptop 2 meters Server 1 meter Google 10 cm Google AI 1 cm

If you get to interview Scarlett, and need somebody to hold the boom, camera, make tea, supply hob nobs, clean shoes…. You have my number…

Add to this the global level of humans who currently and soon will experience "status injury" a sense of no longer feeling useful. Status injury leads to anger - we already have an epidemic of anger. And anger addicts rarely know what make them mad in the first place. Any target will do. And here we are.

Great piece Zoe. "Move fast and break things" has indeed become a poor reflection on today's billionaire techbros that they are quite happy to break things that other people then have to fix. The original ethos was far more confined to their own systems with consequences for them alone (on the whole).

"That report claimed there was currently no reliable way of understanding exactly why AI tools generate the output that they do - even their developers aren’t sure." To me that's the scariest sentence of all. If human developers can't fully grasp what AI is capable of right now, the chances of them being able to do it in the future are incredibly slim.

Thanks Zoe, a fascinating read full of intrigue!

But there's no way that OpenAI would just treat a woman like their own property, literally taking her voice after explicit refusal. They're from Silicon Valley, the tip of the spear in the liberation of humanity from oppression. They support every liberal ideology, and incessantly drone on about their angelic levels of virtue. They would never, right?