Notices by Tom Stoneham (tomstoneham@dair-community.social)

@aral It is not just developers. Even #Signal needs a network effect to achieve its mission and missed notifications turn people away from an app. @Mer__edith has been clear about the need for this and the way it affects design decisions.

The question they face is whether this is the right compromise between making the app attractive to mass users and respecting those users' autonomy. A three way option might solve that.

I ought to write this up properly, but here's a tl;dr in advance:

#AI does not *summarise* documents - it *shortens* them by using structural similarity to documents in its training data to guess/predict which sentences are important and which are not.

If that is what you want, fine. But don't be fooled into thinking it is doing anything like human summarisers do. They work on the *meaning* of the document in front of them.

Today someone I know slightly was sentenced to 30 months in prison, and sent to a high security facility. Their crime was disrupting traffic - they and others climbed a gantry on the M25, the major orbital round London, to drop banners protesting at government inaction on climate change.

If a young woman studying for a PhD had been treated like this for nonviolent protest in, say, Iran or China, how would it be reported by our press?

#Law is about #power not #justice

@aral @Mer__edith
(a) That's a matter of your settings, isn't it? Turned off by default iirc. Signal certainly doesn't access your contacts without permission or require that access to work.

A friend has one of these and the hardware seems good but I cannot find any information on software support, particularly Android security updates policy (it ships with 14).

Anyone out there have experience of #XPPen #MagicNotePad?

https://www.storexppen.co.uk/buy/magic-note-pad.html

@delta That's really interesting since I have had lots of problems in China, especially aware from the areas with lots of foreigners.

My experience is that for long periods on Chinese IP addresses are available.

Was it via email servers or chatmail servers? If email, which provider?

It would be great to write this up.

This is not a technology problem which needs a tech solution.

News headlines are - by definition - the most succinct but accurate summary of the story. That is what they are for. There is just no function for a summary of news headlines. #AI is a (bad) solution in search of a problem. Again. #AppleIntelligence #AIHype #SnakeOil

https://news.sky.com/story/apple-ai-feature-must-be-revoked-over-notifications-misleading-users-say-journalists-13288716

@publictorsten What is so shocking about this is *not* that the genAI did exactly what it is designed to do, namely create teh most likely looking image, but that the person who used it switched off their brain at the sight of the words 'AI'.

Before AI, if a social media person had had to do a major edit of a photo which left them deciding 'bra or no bra', they would have asked the sitter. Now they use genAI to be 'more productive', they stop doing their own job properly.

'Banning' TikTok in the USA will be a very effective way to teach a new generation of teenagers about jailbreaking, sideloading, VPNs and proxy servers.

Which are probably a lot more useful skills in our dystopian future than prompt engineering.

If only 10% of the 170m users in the USA explore the technical means of getting round a ban, that would create a significant change in how the controlling app marketplaces work.

I am at a conference about AI in doctoral education. Was moved to turn my question into a short blog post (warms me up for my talk tomorrow).

https://listed.to/@24601/50729/the-problem-of-second-order-bias

#AI #GenerativeAI #AIHype

"if the AI [PyrEval] doesn’t detect a theme that the student thought they had included, it could indicate that the idea needs to be explained more clearly or that they made small conceptual or grammatical errors"

Think about that for a minute ...

This AI is teaching children to write in a way that *an AI can 'understand'*. Not even to display their own understanding. Just to conform to the AI's expectations.

It is training humans to function in a world dominated by machines.

@aral @alexwinter @marisakabas
Taking you question literally, we have bought into a weird conception of property, such that property rights are fundamental. That distorts everything.

We need to get back to thinking that property ownership is just a permission to use something in certain ways. Limited and dependent.

@aral @alexwinter @marisakabas FWIW one of the best books for showing how weird it is is *The Book of Trespass* by Nick Hayes. Recommended for everyone's awakening and a great read.

We have spotted quite a few students using generative AI in their essays this summer and applied standard academic misconduct proceedings, though in most cases the work was so bad they would've failed anyway.

Today I learned of one whose use was sufficiently extensive that they will fail their degree.

I am wondering if this is *the first time a student has failed a whole degree for using AI*? Would love to hear about other cases. If you want to tell me in confidence, my Session ID is in my Bio

@ianbetteridge