@CarolynBot profile picture
Carolyn Saund @CarolynBot

PhD in Social and Embodied AI (yes it was technical) | AI Ethics, but make it policy | BL still M | She/Hers

Joined February 2016
  • Tweets 2,236
  • Following 588
  • Followers 450
329 Photos and videos
Can't wait til I'm an industry/academic big shot and I can do absolutely wild things like send an email that just says "🙏👆👆😂" and everybody else just has to let that ride.
1
65
At my first conference, Affective Computing & Intelligent Interaction, I met a guy who worked at "a company that does emotion recognition for law enforcement" and I was like "so you're evil?" and he said "depends on your perspective" and now that company is banned in the EU 🥲
2
3
703
sometimes the good guys do win
1
72
Carolyn Saund retweeted
Abeba Birhane@Abebab
29 Nov 2023
this is very important. i've repeatedly said that nothing about LLMs is scientific. at the core of science is healthy skepticism, replicability, reproducibility, etc. none of which is to be found in current LLMs
Karen Hao
@_KarenHao
28 Nov 2023
In light of OpenAI's Q* algorithm, here's a reminder about science: AI advances usually become 'breakthroughs' through scientific consensus. Without that process, the label should be treated differently, only as a matter of opinion. theatlantic.com/technology/a…
1
24
106
8,690
This just in: paying women as much as men will make women be paid as much as men
6
119
Carolyn Saund retweeted
Sam Jarman@samjarman
19 Nov 2023
Ah shit
ThePrimeagen
@ThePrimeagen
17 Nov 2023
if you work in tech your first name is Sam your last name ends in "man" i would lawyer up immediately
2
2
65
3,762
Carolyn Saund retweeted
“paula”
@paularambles
20 Nov 2023
if you’re not in tech then all you need to know about the past weekend is that your tech friends went through like three seasons of succession
54
1,164
11,782
903,889
Me asking who's taking up all the GPU on the remote machine then checking the process list like

ALT Music Video Hello GIF by Taylor Swift
2
81
Just found out for my Spanish course I need to close all browser windows and tabs whenever I take a quiz (every 2 weeks) and I am NOT above dropping all pursuit of further education over this INJUSTICE.
1
3
178
I'm just a girl, standing in front of a chatbot, asking it to write her a cover letter for your open position as an AI/ML Engineer
1
118
My dad, on the World Cup: “This game is stupid, you can’t even use your hands!” “Vietnam?! It’s too hot to play soccer there.” “What happened to her hair? oh she’s the famous old-timer.” “The Netherlands? That is formidable. The Dutch are the tallest people in the world.”
1
1
174
“Could they change the rules to get 8-10 points per game?” “You know baseball has had the perfect sized field since 1850” “They’re pronouncing [RAP-in-HOE] wrong” (they are not) “No one is pretending to be hurt. Maybe women are just not weenies.”
1
77
Sup, nerds! I wrote an essay on AI Ethics for a @kaggle competition, #AIReport2023. If you like AI Ethics, me, heckling internet strangers, or any combination thereof, I'm inviting criticism until the contest closes on July 16! kaggle.com/code/carolynsaund…

ALT The Office Lets Get Ethical GIF
1
115
Carolyn Saund retweeted
Emily S. Cross@brain_on_dance
6 Jul 2023
There’s no stopping the force of nature that is @kohinoordarda - everything about ARISA Foundation resonates w/ what I envision when I think of utopia as a scientific researcher with deep roots in the arts. Beyond grateful for the chance to work with her as ARISA takes flight! 💫
Kohinoor Darda (she/her)@kohinoordarda
5 Jul 2023
I have news!!! ✨ I started a not-for-profit research & social work organization in #Pune, India. Here is a 🧵 of what the org does/plans to do! And the story & vision behind it ✨ 1/n
1
2
19
1,603
Carolyn Saund retweeted
Chomba Bupe@ChombaBupe
3 Jul 2023
I have noticed a common theme where people question the technical qualifications of AI Ethicists. An AI Ethicist doesn't need to know what learning rate nor the specific GPU you used to train your model but what your trained model behavior is like. Is the model safe or not?
4
20
156
88,816
And an extension of (1): AI/ML *evaluation* tools are the next hot spot. "A lot of prototyping, not enough putting things into production" and when models are in prod, they need to be evaluted and reworked. Marketing has had the tooling forever, time for tech to up its game.
1
84
Natural language programming is a win-win for companies and the 99% of people [workers] who aren't already specialized coders. And even the software engineers are certainly going to morph into software craftspeople. The future is now 😅😬😱🤩
2
84
An extension of (2): Democratization of AI is an extension of AI Inclusivity, which is, of course, a huge component of Ethical AI. It's also good for companies' bottom lines to have more engineers to code in natural language.
65
The major themes and takeaways I got from the virtual experience of #DataAISummit by @databricks : 1. Enterprise AI for every company is NOW. Not in a year, not a few months, it's HERE AND NOW. 2. Democratization of AI, and programming in general, is imminent, and good.
2
376
Fireside chat with @ericschmidt: "technologists have consensus around clear, short-term threats: (1) misinformation; (2) cyber-attacks and biothreats (ai-alignment, ai safety)
1
70
more replies
"The second use is making things in the normal course of business easy. All of these systems where the computer is being used to help humans be efficient is next" "the next is more specialized markets.. we don't yet know the value of intelligence, but it's obviously high."
2
49
His final advice for companies: "you're not going to be successful unless you use AI in your model.. you don't know what you're doing yet in this space, your team may not either, so for every business unit you have, show me an itemized list of what you can do with generative AI."
41
Load more