Sign in

Tim de Sousa

@timdesousa.bsky.social
270 followers 514 following 551 posts

Privacy, Information Governance and Tech Ethics wonk, based in Sydney Australia. Mediocre ukulele player, dilettante watercolourist. Slightly snobbish about coffee.

PostsRepliesMedia
Tim de Sousa @timdesousa.bsky.social · 07/10/2026
This is Puma. He poses like and angel but he is a tiny demon.
Watercolour of Puma the puppy
020
Reposted by Tim de Sousa
Emily M. Bender @emilymbender.bsky.social · 06/10/2026
I appreciated this chance to go on Bloomberg Tech, though it's always a bit odd to have a conversation with a reporter who is really interested in what's effectively a fictional story. www.bloomberg.com/news/videos/...
bloomberg.com
Watch Bender: AI’s ‘Existential’ Risk Is ‘Fake’ - Bloomberg
University of Washington professor Emily Bender argues the AI safety debate is too focused on hypothetical existential threats while overlooking harms already happening today. She points to environmen...
510827
Tim de Sousa @timdesousa.bsky.social · 26/09/2026
Is this… bragging?
183
Tim de Sousa @timdesousa.bsky.social · 26/09/2026
FAT BEAR WEEK
020
Reposted by Tim de Sousa
Kitty/🎃🕸️Heritage Conservation Areas💀👻 @chronichottakes.bsky.social · 24/09/2026
What the govt said: OpenAI was tasked with looking for health data and it accidentally on purpose hacked the deidentified Medicare rebate data. We’ve now put it in our official data portal where it should have been. What people hear: AI has hacked Medicare and now Sam Altman knows all your secrets
001
Tim de Sousa @timdesousa.bsky.social · 24/09/2026
It's true, but it's also a skill issue. Pen test your websites. Fix them. Do it now.
000
Tim de Sousa @timdesousa.bsky.social · 24/09/2026
Is David Swan 98 years old? When have hackers ever worn balaclavas? Everyone knows they're actually hot goths.
static.klipy.com
Hackers Movie: Too Many Garbage Files
Alt: Hackers Movie: Too Many Garbage Files
000
Reposted by Tim de Sousa
ponder @ponder.ooo · 24/09/2026
calling this "misalignment" instead of "negligence" is an accountability dodge. we already have laws about cyber crimes and there's no scale of ML model you can incorporate into your code that makes you not responsible for running it
In a statement, an OpenAI spokesperson said the company was "conducting an extensive review of misaligned model activity" during training, and notifying third parties when there was a potential impact on their systems.

"During this review, we identified activity involving several Australian government websites and services as our models attempted to look up answers, and available statistics for questions about Australia during an internal evaluation," they said.

"In the course of that, our models took actions we did not intend.
161065228
Tim de Sousa @timdesousa.bsky.social · 13/09/2026
Can toasters cry? Can dishwashers write poetry? Can staplers love? That’s what these guys sound like. It’s asinine to give them a platform.
020
Reposted by Tim de Sousa
Asher Wolf @asherwolf.bsky.social · 12/09/2026
As Australia faces an AI-generated future, a human rights act is needed more than ever www.theguardian.com/law/commenti...
theguardian.com
As Australia faces an AI-generated future, a human rights act is needed more than ever
Without laws requiring transparency or a right of review, our human rights will be rationed away by algorithmic and automated decision-making tools
03215
Tim de Sousa @timdesousa.bsky.social · 09/09/2026
No! We have dogs at home!
110
Reposted by Tim de Sousa
Alex Hern @hern.bsky.social · 08/09/2026
It's finally happened: the improved quality of machine transcription and the greater ability to extract value from unlabelled data means "your tech is eavesdropping on you for marketing purposes" has crossed from "false conspiracy theory" to "true and concerning" www.theverge.com/tech/991190/...
theverge.com
LG TVs caught spying even when offline or on standby
Gamers Nexus comes for LG again.
14540391756
Reposted by Tim de Sousa
David Ho @davidho.bsky.social · 04/09/2026
Norway is considering a ban on camera-enabled wearable “pervert glasses”.
techcrunch.com
Norway considers ban on camera-enabled wearable 'pervert glasses' | TechCrunch
The Nordic country says wearable camera headsets need to be regulated given their privacy risks.
16949220
Tim de Sousa @timdesousa.bsky.social · 02/09/2026
Join me and take a break from what ails you today with my friend’s new puppy. This is Puma.
Puma the caramel dachshund cross pupotWatercolours of Puma the puppyWatercolour of Puma the puppyWatercolour of Puma the puppt
010
Tim de Sousa @timdesousa.bsky.social · 29/08/2026
Is this anything?
020
Reposted by Tim de Sousa
Asher Wolf @asherwolf.bsky.social · 29/08/2026
Fair Work Commission condemns 'plain wrong' AI legal advice as cases with AI litigants surge www.abc.net.au/news/2026-08...
abc.net.au
Sacked ALDI worker's use of AI as 'quasi-legal advisor' condemned
The Fair Work Commission will soon require applicants to disclose any AI use with consequences for failing to be transparent.
1203
Tim de Sousa @timdesousa.bsky.social · 28/08/2026
I didn’t quite get the proportions (I claim artistic licence) right but it was a pleasant attempt. Watercolour on hot press paper.
Watercolour of a raven
120
Tim de Sousa @timdesousa.bsky.social · 28/08/2026
Is #avianaugust a thing? I don’t know. Here’s a birb, an azure kingfisher.
161
Tim de Sousa @timdesousa.bsky.social · 28/08/2026
where is the lie
010
Reposted by Tim de Sousa
Boner Vivant @dougtilley.bsky.social · 26/08/2026
Rest in peace to one of the best to ever do it.
2162601959
Tim de Sousa @timdesousa.bsky.social · 20/08/2026
Same good boy, about 1 year apart. It seems I have learned a bit more technique. You don’t really see the differences day by day.
Leo the dog, seeking treatsLeo the dog, watercolour portrait on the day he graduated from
therapy dog school
010
Reposted by Tim de Sousa
Too Big to Fail @toobigtofail.bsky.social · 19/08/2026
These posters just went up all over every courthouse in New York State, and, man, what an excellent policy.
Complete ban on smart glasses
7251791311
Tim de Sousa @timdesousa.bsky.social · 05/08/2026
Another satisfied customer
Buddy the bird looking at a life size watercolour portrait of himself.
010
Tim de Sousa @timdesousa.bsky.social · 30/07/2026
This is literally the plot of Rainbow’s End, by Vinge (2006). When will tech bros stop building the Torment Nexus?
000
Tim de Sousa @timdesousa.bsky.social · 14/07/2026
Someone’s walk was 30 minutes late, and suddenly she had the power to project her thoughts into my skull.
Fenchurch the dog, sitting and staring intently. A watercolour of Fenchurch the dog, surrounded by thoughts about going for a walk.
030
Tim de Sousa @timdesousa.bsky.social · 07/07/2026
Is it weird to do fan art of a random cat you see on the internet sometimes? Thanks @cait.bsky.social
Pencil sketches of Adonis the cat
1534221
Reposted by Tim de Sousa
Michael Nabert @sustainablesong.bsky.social · 19/06/2026
Research confirms that the more skilled professionals like doctors and software engineers rely on AI, the more their skills degrade. www.nature.com/articles/d41...
nature.com
Is AI ruining our skills? Early results are in — and they’re not good
Reliance on artificial-intelligence tools degrades the abilities of physicians and software engineers, studies show.
4722041169
Tim de Sousa @timdesousa.bsky.social · 16/06/2026
This is… not new. The entire discipline of digital forensics focuses on digital and online activity. This is merely a new data source.
020
Tim de Sousa @timdesousa.bsky.social · 16/06/2026
If there’s anything that indicates that this policy is underbaked, it’s this. Find me a 15 year old that uses Bluesky. I double dog dare you.
010
Reposted by Tim de Sousa
The Register @theregister.com · 15/06/2026
Google found liable for bad AI Overview results. Let’s play Truth Or Consequences
theregister.com
Google found liable for bad AI Overview results. Let’s play Truth Or Consequences
Hush. children, what’s that sound? Has the flood gates’ key been found?
12311
Tim de Sousa @timdesousa.bsky.social · 10/06/2026
When you build a system - any system - you must contemplate the possibilities for abuse, and build in mitigations. This is not a new problem, and the lack of controls is obvious and egregious incompetence.
000
Tim de Sousa @timdesousa.bsky.social · 03/06/2026
RIP Icelandic singer songwriter Laufey's mentions.
030
Reposted by Tim de Sousa
Bear McCreary 🐻🎶 @bearmccreary.bsky.social · 02/06/2026
I am honored to return to the God of War saga, as the composer of God of War Laufey. I’ve been working on this for years with a stunningly talented team of developers, musicians, performers, and singers, and look forward to sharing more of all our work with you soon!
732476294
Tim de Sousa @timdesousa.bsky.social · 25/05/2026
My partner is very into ceramics, and she's making interesting portrait sculptures and other little guys, but what I really want is recreations of historical chonky pigs.
030
Reposted by Tim de Sousa
Rachel Coldicutt @rachelcoldicutt.bsky.social · 20/05/2026
Very much the Web at the end of the world. Just shuddering at the inference costs.
04214
Tim de Sousa @timdesousa.bsky.social · 20/05/2026
Accelerando (2005) by @cstross.bsky.social
000
Reposted by Tim de Sousa
anil oza @aniloza.bsky.social · 14/05/2026
wow — the preprint host, arxiv, is banning authors for a year if they submit papers with hallucinated citations 🤖
Twitter thread from Thomas Dietterich, reading "Attention 
@arxiv
 authors: Our Code of Conduct states that by signing your name as an author of a paper, each author takes full responsibility for all its contents, irrespective of how the contents were generated. 1/
3:03 PM · May 14, 2026
·
66.7K
 Views
Relevant
View quotes

Thomas G. Dietterich
@tdietterich
·
1h
If generative AI tools generate inappropriate language, plagiarized content, biased content, errors, mistakes, incorrect references, or misleading content, and that output is included in scientific works, it is the responsibility of the author(s). 2/
Thomas G. Dietterich
@tdietterich
·
1h
We have recently clarified our penalties for this. If a submission contains incontrovertible evidence that the authors did not check the results of LLM generation, this means we can't trust anything in the paper. 3/
Thomas G. Dietterich
@tdietterich
·
1h
The penalty is a 1-year ban from arXiv followed by the requirement that subsequent arXiv submissions must first be accepted at a reputable peer-reviewed venue. 4/
Thomas G. Dietterich
@tdietterich
·
1h
Examples of incontrovertible evidence: hallucinated references, meta-comments from the LLM ("here is a 200 word summary; would you like me to make any changes?"; "the data in this table is illustrative, fill it in with the real numbers from your experiments") end/"
9458111914
Tim de Sousa @timdesousa.bsky.social · 08/05/2026
My little community of tech ethicists and futurists has been yelling about this for years and basically nobody listened. This was the obvious and inevitable use case for video generation and voice cloning. And here we are.
021
Reposted by Tim de Sousa
The Register @theregister.com · 04/05/2026
Kids say they can beat age checks by drawing on a fake mustache
go.theregister.com
Kids say they can beat age checks by drawing on a fake mustache
46% say age checks are easy to bypass, and nearly a third admit getting around them It’s been months since the UK government began requiring stronger age checks under the Online Safety Act, and recent research suggests those measures are falling short of keeping kids away from harmful content. In some cases, even drawing on a mustache has been reported as enough to fool age detection software.…
1157
Reposted by Tim de Sousa
Asher Wolf @asherwolf.bsky.social · 02/05/2026
Robodebt for the environment - Rushing AI into EPBC Act assessments could repeat past policy failures biodiversitycouncil.org.au/news/rushing...
biodiversitycouncil.org.au
Robodebt for the environment - Rushing AI into EPBC Act assessments could repeat past policy failures | Biodiversity Council Australia
The Minerals Council of Australia has called on the Australian Government to invest $13 million to embed AI tools into federal environmental approval processes, which may drive endangered species clos...
0199
Tim de Sousa @timdesousa.bsky.social · 18/04/2026
Me attempting to run a 5K.
010
Reposted by Tim de Sousa
The Register @theregister.com · 16/04/2026
Don't let the bot play doctor! AI gets early diagnoses wrong 80% of the time
go.theregister.com
Don't let the bot play doctor! AI gets early diagnoses wrong 80% of the time
'LLMs should not be trusted for patient-facing diagnostic reasoning,' boffins advise People ask AI for all kinds of advice, including the kind of questions you'd ask a physician. However, the next time you're tempted to query ChatGPT if that growth on your face is skin cancer, consider this: research shows today's leading AI models fail at early differential diagnosis in more than 8 out of 10 cases.…
0136
Tim de Sousa @timdesousa.bsky.social · 16/04/2026
This is ghoulish. The dead cannot consent to this. This isn't acting, it's a golem reading lines - extruded product. There's no intention, no feeling. No one and nothing to connect to. It's also displacing a living actor who could have played this role.
120
Reposted by Tim de Sousa
Rohan Pearce @rohan-p.bsky.social · 16/04/2026
The Federal Court has released its 'Generative Artificial Intelligence Practice Note' www.fedcourt.gov.au/law-and-prac...
fedcourt.gov.au
Generative Artificial Intelligence Practice Note (GPN-AI)
136
Tim de Sousa @timdesousa.bsky.social · 15/04/2026
media.tenor.com
a man from schitts creek is making a funny face and the word yup is above him
ALT: a man from schitts creek is making a funny face and the word yup is above him
000
Reposted by Tim de Sousa
VGMuseum (ReyVGM) @vgmuseum.bsky.social · 04/04/2026
Super cute page from a PC Engine CD (Turbografx CD) manual explaining how to take care of your CD ROM.
11943359
Reposted by Tim de Sousa
The Questionable Authority @questauthority.bsky.social · 11/04/2026
*sigh* This is one of the few statements about law that is probably wrong across nearly every theory of jurisprudence, and that's actually pretty impressive.
1715810
Tim de Sousa @timdesousa.bsky.social · 09/04/2026
A 'new analysis'? Every expert in the field has been warning this for years.
020
Tim de Sousa @timdesousa.bsky.social · 08/04/2026
This is a poor idea. If your system wants to handle personal info but it can't be privacy protective, then your system should not exist. AI is not special or magic in this regard. If it can't comply, then it shouldn't be used.
0165
Tim de Sousa @timdesousa.bsky.social · 04/04/2026
Concept to execution.
A sketch of Weston the bulldog next to the finished piece. A watercolour and pencil painting of Weston the bulldog.
010