Reading this comparison between work on "AI scheming" and oldschool chimp+language research and just... disappointed by how invested so many people are in seeing modern day ELIZA as a reasoning machine #NowReading
Computational linguist specializing in #NaturalLanguageGeneration with an interest in linguistic complexity.
Advanced Research Fellow at #UniversityOfAberdeen, living in #Edinburgh
Publishing under David M. Howcroft
More personal at @_dmh
Reading this comparison between work on "AI scheming" and oldschool chimp+language research and just... disappointed by how invested so many people are in seeing modern day ELIZA as a reasoning machine #NowReading
If you have a parent who was born in Canada, it's very easy to get certified as a Canadian citizen, since it's verifying that "you were one all along". Here's more info: https://www.canada.ca/en/immigration-refugees-citizenship/services/canadian-citizenship/proof-citizenship.html
But we are in an unusual position where if you have *any* born-in-Canada ancestors, you can probably be certified as a Canadian citizen. This won't be true forever, so don't wait!
If you are trans, write a personal letter saying so and asking it to be expedited.
If you've ever thought "It's such a bummer that farm jobs suck so bad, no American would ever wanna do them,"
NONE OF THAT IS TRUE.
Here's what building a proper damn farm sector with real jobs & upward mobility would look like. (Extremely abridged bc 8 minutes.)
Also, it's really weird how this conversation in the US always revolves around "Making farm jobs good enough for Americans."
Farm jobs should be good jobs no matter who's doing them.
Even a lot of self-described progressives seem to have a hard time getting that!
knock it off it's weird
You work with the staff you have.
Yooooo! #INLG2025 is hosting a workshop on "LLM Reasoning on Medicine"! Topics of particular interest to me:
- Eval methods...for medical NLG
- Mitigation strats for hallucination
- Safety+ethical considerations in deploying LLMs in healthcare
- many others...
https://2025.inlgmeeting.org/workshops-tutorials-llm-medicine.html
Regarding the *chatgpt unsurprisingly gives misogynist advice* story, I'm thinking about the fact that these tools prioritise chat interaction, which isn't even an effective model for the things they can do well imo
Presumably this choice is a reflection of the desire to replace human workers and to gatekeep access to information / the web, but it's a huge component in the misunderstanding of what they do I think
https://thenextweb.com/news/chatgpt-advises-women-to-ask-for-lower-salaries-finds-new-study
Countdown clock for the #INLG2025 direct submission deadline: https://www.timeanddate.com/countdown/generic?iso=20250718T235959&p0=3399&msg=INLG+2025+Submission+Deadline+%28direct+submissions%29&font=sanserif
Submit to the longest running conference focused on #NaturalLanguageGeneration :D
(ARR commitment deadline is August 7th)
Defending science in public we often talk about 'peer reviewed science'. But could this framing contribute to undermining trust in science and holding us back from improving the scientific process? How about instead we talk about the work that has received the most thorough and transparent scrutiny?
Peer review goes a step towards this in having a couple of people scrutinise the work, but there are limits on how thorough it can be and in most journals it's not transparent. Switching the framing to transparent scrutiny allows us to experiment with other models with a path to improvement.
For example, making review open to all, ongoing, and all reviews published improves this. When authors make their raw data and code open, it improves this.
It also gives us a way to criticise problematic organisations that formally do peer review but add little value (e.g. predatory journals). If their reviews are not open and observably of poor quality, then they are less 'thoroughly transparent'.
So with this framing the existence of 'peer reviewed' but clearly poor quality work doesn't undermine trust in science as a whole because we don't pin our meaning and value on an exploitable binary measure of 'peer reviewed'.
It also offers a hopeful way forward because it shows us how we can improve, and every step towards this becomes meaningful. If all we have is binary 'peer reviewed' or not, why spend more effort doing it better?
In summary, I think this new framing would be better for science, both in terms of the public perception of it, and for us as scientists.
Reminder that the 'bystander effect' isn't real, Kitty Genovese's neighbours didn't ignore her attack (most didn't even witness it, two people called the cops, someone even went out and held her while the ambulance arrived), and a lot of psychology that's really well known from this period is just total bunk.
I was taught this uncritically when I was studying psychology and it'll take a generation or more for it to really start getting the critical examination it needs.
The first papers are already submitted, but for anyone hoping for an extra week to work on their paper, I have very good news!
The deadline for submitting to #INLG2025 has been extended until July 18th (end of the day anywhere on earth, AoE).
Submit your work on generation!
So. I've been on field manual labor crews where most of my coworkers were convicts.
I've also worked with a lot of tech companies on automating farms.
And this press conference is incredible. Rollins hasn't the foggiest clue what she's on about.
Every farm job that CAN be automated, already is. Let's start there.
She thinks... nobody's ever tried to automate picking fruit? Really?
And then we'll talk about the "tehe we'll just make the Medicaid people work the farms" part.
Announcing 4 keynote speakers for #INLG2025!
Verena Rieser (Google DeepMind)
Hadas Kotek (Apple)
Minlie Huang (黄民烈; Tsinghua University)
Mike White (Ohio State University)
It's shaping up to be a great conference!
Don't forget to submit by 14 July AoE:
https://2025.inlgmeeting.org/
Public transport is no longer available to all public if the only way to buy a ticket to ride physical train, tram or bus is with:
1. a credit or bank card (issued by a financial organisation)
2. an app created by the train transport business that has to be downloaded with paid mobile service & run on your personally purchased smartphone that has to be compatible.
You cannot pay with “cold, hard cash”💰, invented for service to value exchange.
“Exclusionary convenience” has become mandatory!
Maybe we should start complimenting AI art with "Wow, that is *really* statistically plausible!"
On Tuesday 17th June 2025 MSPs votit unanimously tae pass the Scottish Languages Bill intae law.
This'll be the first time oor language is recognised in Scots law!
It's the haurd wark o a guid wheen o fowk, wi the backin o a hunner year o activism wi thoosans involved, an millions o Scots speakers, fae the 2,444,659 thit reportit Scots skills in the 2022 Scotland's Census gaun back throu aw thae speikers ower the past near millenium.
...
Presented #ASICA at #GramConf2025 today :)
Ahah, the actual conference hashtag is #GramConf2025!