Tom Stoneham

https://dair-community.social/@tomstoneham

Academic philosopher at a UK university.

Likes: #Linux, #GrapheneOS and #e2ee

Cares about: #DataPrivacy and #ProportionalRepresentation

Learning: 🇩🇪

Currently thinking about value alignment, data privacy and the digital divide.

Would like to ground ethical critique of tech in political consensus at local levels (as opposed to global approaches like human rights).

Motto: Think slowly and fix things
Blog: https://listed.to/@24601
Contact: https://listed.to/p/764ujhoNas
Study with me: https://www.york.ac.uk/study/postgraduate-taught/courses/ma-applied-ethics-governance-data-privacy/

Public Key

npub1ttdjsgp49cyv72ftyctqlr8uf4dpg9cs5fg3gk4p7xexmn29s99q35tvc8

NIP-05 Address

[email protected]

Profile Code

nprofile1qqs94kegyq6juzx09y4jv9s03n7y6ks5zug2y5g5t2slrvnde4zczjspz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dsuss22a

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1ttdjsgp49cyv72ftyctqlr8uf4dpg9cs5fg3gk4p7xexmn29s99q35tvc8

Show more details

Published at

2024-06-26T21:34:53+02:00

Kind type

0 Metadata

Event JSON

{ "id": "42c7d294ef9b25e41b1ca6439f0104514bef2744b74b14a67ebb0560197c4a4b", "pubkey": "5adb2820352e08cf292b26160f8cfc4d5a141710a251145aa1f1b26dcd45814a", "created_at": 1719430493, "kind": 0, "tags": [ [ "t", "linux" ], [ "t", "DataPrivacy" ], [ "t", "e2ee" ], [ "t", "proportionalrepresentation" ], [ "t", "grapheneos" ], [ "proxy", "https://dair-community.social/users/tomstoneham", "activitypub" ], [ "L", "pink.momostr" ], [ "l", "pink.momostr.activitypub:https://dair-community.social/users/tomstoneham", "pink.momostr" ], [ "expiration", "1722022493" ] ], "content": "{\"name\":\"Tom Stoneham\",\"about\":\"Academic philosopher at a UK university. \\n\\nLikes: #Linux, #GrapheneOS and #e2ee\\n\\nCares about: #DataPrivacy and #ProportionalRepresentation\\n\\nLearning: 🇩🇪 \\n\\nCurrently thinking about value alignment, data privacy and the digital divide. \\n\\nWould like to ground ethical critique of tech in political consensus at local levels (as opposed to global approaches like human rights).\\n\\nMotto: Think slowly and fix things\\nBlog: https://listed.to/@24601\\nContact: https://listed.to/p/764ujhoNas\\nStudy with me: https://www.york.ac.uk/study/postgraduate-taught/courses/ma-applied-ethics-governance-data-privacy/\\n\",\"website\":\"https://dair-community.social/@tomstoneham\",\"picture\":\"https://cdn.masto.host/daircommunitysocial/accounts/avatars/109/304/508/297/555/813/original/41f80fefc20a701f.png\",\"banner\":\"https://cdn.masto.host/daircommunitysocial/accounts/headers/109/304/508/297/555/813/original/b0abdb8fc48b7ed9.png\",\"nip05\":\"[email protected]\"}", "sig": "72d7474b8a740624449aa1a7f13ca902e089a322f9aaadf810dc0819c2a18fdcf19bfe82a11ecca4e4edc51f6154931962de0f1128c98f9a1fe929054ee5935d" }

Last Notes

2024-07-20 10:36:36

- reply

What's a *young* female name? 😆

2024-07-14 17:27:14

- reply

Potato lattices. You can buy them under that name in Tesco. Also sometimes lattice fries in fast food outlets.
https://www.tesco.com/groceries/en-GB/products/286999595

2024-07-12 10:11:23

For sale:
Fairphone 4 with /e/ OS. Excellent condition.
£275 plus postage UK/EU
#murena #eos2 #fairphone #Fairphone4

2024-06-25 15:21:08

This morning I was moved to have a rant about AI Teaching Agents.
#AI #AIEthics #JustStopAI
https://listed.to/@24601/52636/44-million-teachers

2024-06-18 18:00:25

I keep looking at this headline from the Financial Times (17 April) and puzzling how these smart economists could be taken in.
The 'demand' is being artificially created by the suppliers bundling the product with just about everything digital and hiding its true costs.
"Booming AI supply threatens global electricity supply" has an obvious and much simpler solution: #JustStopAI
https://cdn.masto.host/daircommunitysocial/media_attachments/files/112/638/432/143/257/137/original/8980c6d497d18f26.png

2024-06-11 18:37:39

- reply

Good point re PPE 😆 (though I did PPE at Oxford during the time it was breeding Prime Ministers ...)
But this isn't even an ethical error. For a philosophy undergrad it is basic methodology to distinguish belief from truth. Standard critiques of verificationism are taught the world over.

2024-06-11 18:34:57

- reply

To be fair, they were trying to find a technical solution to having disagreement in the (labelling of) training data, but it boiled down to something which seemed to 'blend' credence with degrees of truth.

2024-06-11 18:05:20

- reply

I didn't want to make this about an individual so much as the culture at Google Research. But I believe the webinar recording will be available soon.

2024-06-11 17:50:07

In a webinar where a Google Research responsible AI person (who can remain anonymous) has just said:
"I do not think there is a ground truth because even for simple things people will disagree."
Never in human history has there been such a need for more philosophers and fewer scientists. Second year undergraduates should be able to critique this conflation of epistemology and metaphysics.
#AI #generativeAI #Philosophy

2024-06-09 16:44:31

- reply

On relfection, I guess it is insurance against serious misuse of the system: phone numbers are almost always tied to real world identities.
(I think the UK may be one of the easier places to avoid that.)

2024-06-09 13:48:08

I am struggling to think of the security need for an AI chatbot to want phone number verification. It is already linked to my Telegram account. What would my phone number add that benefits me rather than them? I assume they are going to verify by SMS, but *what* are they verifying other than that my Telegram account is linked to a real number?
#AI #genAI #Copilot #Telegram
https://cdn.masto.host/daircommunitysocial/media_attachments/files/112/586/483/391/932/577/original/7939d5a1f01c0435.png

2024-06-04 22:37:16

Maybe I am just feeling tired and cynical, but it seems to me that far too many, maybe most, academics are just interested in *what other academics say*.
They build on what other academics say, they criticise what other academics say, the engage with what other academics say.
No attempt to step outside the bubble and find the ground truth.
No wonder academia is so vulnerable to #LLM fakery.

2024-06-04 12:16:38

- reply

The Chinese messaging app #WeChat already has #ChatControl. I know because I have been banned from uploading images etc for attempting to upload this one in November 2022 during the #WhitePaper protests (expressing solidarity).
I haven't tried to travel to China since, but will be applying for a visa later this year and that will be the test of quite how pernicious this approach is.
https://cdn.masto.host/daircommunitysocial/media_attachments/files/112/557/819/338/901/904/original/062521bf173f7034.jpg

2024-05-30 20:37:03

#JustSayin #HeatStrike Watch this space:
https://heatstrike.uk/

2024-05-30 19:46:00

I am struggling to understand why Zoom on Linux (Ubuntu 22) is so flakey* for me over WiFi when same 100mb/s WiFi set-up works fine for Zoom on other devices and for streaming.
Same behaviour on two Linux devices and three WiFi networks.
Switching to LAN wired connection seems to solve it.
*Freezes. Audio drops. Crashes when picture-in-picture.

2024-05-28 16:22:25

Who are you to judge me #Discord? I was reading!
https://cdn.masto.host/daircommunitysocial/media_attachments/files/112/519/153/206/552/131/original/39a28b74c3af0696.png

2024-05-28 15:01:46

- reply

But that aside, there is also something really important underlying your intuitions here.
North American privacy scholars have for over a century been promoting the idea that privacy is ultimately a property right.
I think this is a HUGE mistake, partly because it presupposes a very culturally specific notion of property that we can blame Locke for, but mainly because it leads to obvious bad consequences in the internet age, like scraping social media to train AI. (2/2)

2024-05-28 14:58:27

- reply

Thank-you for taking the time to read and respond! It is greatly appreciated.
Speaking as a philosopher, I would say that ethics - unlike perhaps law and regulation, though that may be contested - is and has to be fuzzy otherwise it undermines its own relevance. Humans are messy and complicated and ethics should reflect this or it can no longer be a guide to behaviour for real people. Rule are for robots.
(1/2)

2024-05-28 14:03:38

- reply

I assume this is not 'too big to fail' like banks but the legal situation being that - since this is an essential service - if the company fails it goes directly into public ownership (as happened with LNER) rather than bankruptcy. And whoever ends up owning the company - however 'bankrupt' - is still liable for its debts.
Otherwise those creditors would be able to do what they wanted in a bankruptcy fire sale and £16m people would be without water.

2024-05-28 12:47:39

Can someone better at economics explain to me why Thames Water's £15bn debt is such a barrier to renationalisation?
Since TW (or any water company) cannot go bankrupt in the normal way, that debt is *already* a government liability.
So surely it is just a *choice* about how we do public accounts not to include it as part of 'government debt'? Do public accounts not include properly audited liabilities?

2024-05-28 12:15:09

This is really useful. Please share widely.
https://voteclimate.uk/

2024-05-28 12:11:42

- reply

There is such an obvious inconsistency between trickle *down* and *rising* tide that I am endlessly surprised anyone gets away with combining these metaphors.
*Rising tide* is actually a very good metaphor for the economics of redistributive taxation, since the beneficiaries will spend a greater proportion (often >100%) of their income thereby creating growth.

2024-05-28 07:38:52

My longer answer to a fascinating question asked by @npub1huw…c5x7 about the data on old computers.
#Data #DataPrivacy #Ethics
https://listed.to/@24601/51997/privacy-and-old-data

2024-05-27 07:51:08

- reply

Surely you both already knew that Zoom/Meet or whatever was likely using your calls to train AI?
The real worry about #Recall is that it sidesteps any effort to use secure #e2ee communications by recording the plaintext.
It is just delayed client-side scanning and a similar threat to our right to private correspondence.

2024-05-26 21:39:43

- reply

I am going to need to think about this a bit more, but I am predicting my conclusion will be that:
We all know and expect to leave 'archive' traces of our lives (photo albums and diaries end up in car boor sales).
But we trust others to respect us in how they use that information.
We can buy old photo albums to laugh at people, or to better understand their lives and times. That is the real ethical choice.
So look but be respectful.

2024-05-26 11:40:31

- reply

You can see how it appealed to some policy wonk.
Apart from attracting the gammon vote, it bulks up military numbers and spend by introducing the equivalent of PCSOs. And also funnels some young voters away from HE and the development of critical thinking into a culture of discipline and blind obedience.
Even if entirely voluntary, it might serve a Tory purpose.

2024-05-26 09:42:50

- reply

He probably hasn't realised that women aren't a minority group. Like umbrellas and contactless payments, this is all outside his lived experience.

2024-05-26 09:00:58

This is a typical early campaign trial balloon of a half-baked idea before the manifestoes are finalised.
What is really surprising is how it is to be funded: "by cracking down on tax avoidance and evasion."
Tories wanting their supporters to pay more tax without 'raising taxes'?
As soon as CCHQ put a number on what they will raise by that route, every other party can spend it in their manifesto with impunity.
https://news.sky.com/story/sunak-says-he-will-bring-back-national-service-if-tories-win-general-election-13143184

2024-05-25 08:09:06

- reply

Solving human/social problems by banning things doesn't have a history of success.
Anyone remember John Major and 'dangerous dogs'?
More magical thinking, this time #ReguloSolutionism

2024-05-24 12:30:36

- reply

Indeed, the Romans placed a very high value on providing certain luxuries throughout their empire, but only for the colonial power, not the colonised.
It is worth reflecting on how many people on earth *still* don't have access to decent plumbing (and why).

2024-05-24 11:02:58

"Nearly all of human life, then and now, takes place far from hot baths."
Simone Weil, _The Iliad, or the Poem of Force_

2024-05-23 09:01:03

Just checking out an NCP carpark for today and noticed that the hourly rate is *£1/h* less if you pay by their app.
That cannot all be reduced transaction costs. It must be in large part what they get from selling your data.
#DataPrivacy #SurveillanceCapitalism

2024-05-23 07:56:43

Why does it always rain on me?
Was it because 'We prioritised energy security and your family finances over environmental dogma in our approach to Net Zero'?

2024-05-23 07:50:56

- reply

"We prioritised energy security and your family finances over environmental dogma in our approach to Net Zero."
Many parts of the UK saw a month's average rainfall in a day yesterday. But Rish! is proud of approving more fossil fuel extraction.

2024-05-22 19:45:38

Even when the sun is shining I can't avoid the lightning

2024-05-22 19:28:35

Is it because I lied when I was seventeen?

2024-05-22 18:06:13

@npub1cqh…8tnn
From the URL it looks like @npub168d…tu7v originally called it a 'ban'. It isn't a ban. It is a perfectly normal collaborative solution to a collective action problem.
These ways of solving social problems have been around as long as humans and a lot longer than laws. They work.
(Yes, there will always be a few freeriders, but there are always people who break laws as well.)
https://www.theguardian.com/technology/article/2024/may/22/st-albans-headteachers-call-for-under-14s-smartphone-ban

2024-05-22 14:29:40

- reply

'Information that is publicly available is thereby no longer private' is a very deep-seated idea.
It ultimately comes from US legal (and philosophical) theory seeing privacy as ultimately a property right.
'Other conceptions of privacy are available' (and should be robustly defended)

2024-05-22 11:55:34

- reply

I am puzzled why everyone is reaching for a ban. It seems the wrong tool to solve a complex social problem.
What is really being banned is *parents* buying smartphones for their children (or otherwise allowing it).
Some do it because they think the benefits outweigh the risk of harm. Others just cave in to pressure. Yet others don't understand the risks of harm.
A ban outlaws the good parents and treats the weak/bad parents as incorrigible failures.
#ReguloSolutionism will fail

2024-05-22 09:27:29

"Yet again, LLMs show us that many of our tests for cognitive capacities are merely tracking proxies."
Some thoughts on genAI 'passing' theory of mind tests.
#AI #generativeAI
https://listed.to/@24601/51831/minds-and-theories-of-mind

2024-05-21 22:33:20

Feeling very smug that when my employer announced the schedule for Windows 11 roll out, I dug deep in the policies and found I could opt for Linux installed on my laptop.
Sent it in quick before they changed their minds ...

2024-05-21 21:23:17

- reply

Or the surveillance desires of companies bigger and more powerful than most states?

2024-05-21 18:31:08

- reply

What can we make of these reports?
I am reminded of @npub1fdr…lvhs pointing out that what Google/Meta etc are *best* at is persuading advertisers that the targeted ad services are effective. Actually producing effective advertising channels is secondary to that.

2024-05-21 13:44:35

- reply

Actually, it is interesting to look at that specific claim and its ambiguities:
"we have found 10 areas in the software development life cycle where GenAI can boost productivity and speed by 20% to 50% right now"
10 areas, 2 dimensions and a massive spread. Does that mean you can get at least 20% and up to 50% boost in both productivity and speed in all 10 areas? Or that the maximum boost seen in any area on either dimension is 50%?

2024-05-21 13:34:59

I really enjoy #AcademicProcrastination
By and large it involves getting distracted reading things tangentially relevant to something you have already done or something that isn't due for months.
It is very relaxing and a lot more fun than the thing due next week.

2024-05-21 13:28:28

- reply

That's how it starts ...
https://www.imdb.com/title/tt0319262/

2024-05-21 11:22:55

- reply

These tech companies really are like teenagers trying it on to see what they can get away with.
Trouble is, they are teenagers who are bigger* than their parents and legal threats mean little to them.
*i.e. have enough cash to shrug off any legal process including fines.
We need an alternative to #ReguloSolutionism ASAP

2024-05-21 11:16:55

- reply

Thx for that link (PWC website is horrendous to find stuff on - I am still not sure what @npub168d…tu7v was reporting on).
This one is optimistic though those error bars are what we should be looking at!
https://www.pwc.com/gx/en/issues/technology/path-to-generative-ai-value.html

2024-05-21 10:29:35

- reply

Here is one from yesterday. No link to AISI let alone the report you are referring to. Which, it turns out, is much more wide-ranging and interesting than the small part of it you report upon for clickbait.
Sloppy journalism.
https://www.theguardian.com/technology/article/2024/may/20/ai-chatbots-safeguards-can-be-easily-bypassed-say-uk-researchers

2024-05-21 10:27:32

ANyone else getting fed up with @npub168d…tu7v refusing to link to the sources for their articles (unless they are on X ffs!)?
Take this one, which reports on a PWC report. There isn't even a link to PWC (just the internal Guardian page of their stories about PWC), let alone to the actual report.
It is this approach which places a barrier to critical evaluation of reporting and encourages #FakeNews.
https://www.theguardian.com/technology/article/2024/may/21/productivity-soars-in-sectors-of-global-economy-most-exposed-to-ai-says-report

2024-05-20 22:46:51

- reply

I have logged in to hotel WiFi the world over with [email protected]
(I suspect it is anyone's for obvious Google censorship reasons)

2024-05-20 22:42:14

- reply

Apple implemented client-side scanning and then had to roll back. Google is doing it by shifting the baseline. I'm sure Microsoft are working on a rollout too.
Why? Why are these companies so keen on the technology?
Because they are effectively (I am prepared to argue literally) states and their users are their citizens and all states like to monitor and control their citizens as much as they can get away with.

2024-05-19 19:50:47

- reply

'Still the way' ????!!!!!
WTAF

2024-05-17 10:11:33

Fuck off. Now. And don't come back. (But please pay your taxes on the way out.)
https://www.theguardian.com/politics/article/2024/may/17/rishi-sunak-and-akshata-murty-rise-up-rich-list-as-fortune-increases-by-120m-pounds

2024-05-15 20:22:05

- reply

Didn't someone write a play called "Six AIs in search of a use case"?

2024-05-15 20:19:21

- reply

I'm not apologising. Never apologise! Je ne regrette rien.

2024-05-15 20:09:18

On a train from a long afternoon protesting at the British Insurance Brokers Association (Only Fools Insure Fossil Fuels) refreshing myself with what M&S chooses to call a 'Mojito'.
True pirates and historians of alcohol will know it is really El Draco. Invented by Sir Francis Drake, possibly to get more vitamins into his crew.
Which is of course the only reason I am having one now.

2024-05-15 20:00:57

Stop #EACOP and #InsureOurFuture
#RedRebels at the #BIBA conference in Manchester today.
https://cdn.masto.host/daircommunitysocial/media_attachments/files/112/446/409/929/782/369/original/30688f5f26562df0.jpg

2024-05-15 10:26:07

I don't use genAI/LLMs for all the reasons.
But at last I have found an ML tool which actually enhances my life: whoBIRD
This does realtime identification of bird songs. It is useful, effective and works entirely on the phone (including offline) so no nasty data centres.
#AI #AIEthics
https://github.com/woheller69/whoBIRD

2024-05-12 14:57:35

- reply

Reverse image search?

2024-05-12 14:57:11

- reply

By and large, history is on the side of civil disobedience and echoes with names like Rosa Parks and Gandhi.
Why don't politicians get that if significant numbers of people are prepared to break the law *and take the consequences* in order to protest against the government, then inventing new crimes they have committed is obviously authoritarian?
They have so much faith in political institutions that they cannot distinguish procedural from substantive justice.