Katherine Bailey (@katherinebailey) header
Katherine Bailey (@katherinebailey) profile picture

SW Eng @Shopify. I no longer work in AI because I'm so over it. Diver, amateur classical musician, lover of languages.

Toronto, Canada
Joined September 2008
Photo by Katherine Bailey (@katherinebailey)Photo by Katherine Bailey (@katherinebailey)Photo by Katherine Bailey (@katherinebailey)Photo by Katherine Bailey (@katherinebailey)Photo by Katherine Bailey (@katherinebailey)Photo by Katherine Bailey (@katherinebailey)Photo by Katherine Bailey (@katherinebailey)Photo by Katherine Bailey (@katherinebailey)Photo by Katherine Bailey (@katherinebailey)
Is there a support group for people who have a more cynical view of LLMs than all of their coworkers? Asking for a friend.
5
1
15
1,635
My optimistic take on where we’re at right now: this is the pre-dawn of the post-bullshit age.
7
1,070
An idiot wind is blowing
1
4
1,339
Katherine Bailey retweeted
It just occurred to me that there's been a big plot twist: 2016: Huggingface - we're going to make a chatbot OpenAI - we're going to do open AI 2022: Huggingface - we're making AI open OpenAI - we made a chatbot ... 😅
12
165
15
2,001
Given that most AI researchers don’t even try to present a philosophically sound definition of intelligence, this is a laughably far-fetched claim.
The case for ambition in artificial intelligence research: Within your lifetime, AI researchers will understand the principles of intelligence—what it is and how it works—well enough to create beings of far greater intelligence than current humans.
2
5
Shockingly flawed reasoning about the relationship between evolutionary constraints and intelligence (it is literally the opposite of what’s expressed here). The field of AI is becoming more of a joke every day.
Replying to @togelius
A system that can learn, generalize, and reason like humans by building models of the world. It will be more general and more powerful than humans because arbitrary evolutionary constraints are removed, but still limited because it’s not some mythical “universal intelligence”.
1
3
“There are typically around 1,000 pieces of sound in a cumulonimbus cloud.” GPT3 wisdom at its finest. economist.com/by-invitation/…
1
I get tons of these too and I’m just a Staff Software Developer. They sure are desperate to sell all those ML solutions.
You’re unable to view this Post because this account owner limits who can view their Posts. Learn more
1
Just booked a flight to Vancouver for a team get together next month - it'll be my first time back in 8 years! Might have to make a quick trip out to Bowen Island - of all the places we've lived, nothing has really come close to that paradise.
3
1
8
People who think evolution is an optimization process need to read a bit more about evolution.
People who think evolution works through random mutations and selection need to explain how intelligent life appeared using nothing else. Clearly, any optimization process is more efficient if it uses some sort of gradient estimation.
3
So um... anyone else ever move from software engineering to data science and back to software engineering again? 🙃
3
9
It’s World Dictionary Day! To mark the occasion, here’s an A to Z thread of @HaggardHawks facts about the 26 letters of our alphabet... 🧵
24
571
185
1,242
Banjo swimming in Lake Ontario
6
Wow. This actually makes David Chalmers look like a fool.
This Post is from an account that no longer exists. Learn more
2
2 years to the day since moving to Toronto, we finally went to see Niagara Falls. Nice to see something so grandiose. Haven’t left Ontario since Jan 2020 and it’s a province whose charms are often quite... subtle.
2
8
I participated in a philosophy workshop earlier this week - that’s a first! It was on Feminism, Social Justice, and AI. Got great feedback on our paper The Bias Dilemma: Reframing the Debate About Algorithmic Bias, and enjoyed discussions around the many other interesting papers.
5
A few years ago, this same thing was being illustrated with Turkish (o bir doktor, o bir hemşire => he is a doctor, she is a nurse). Then google claimed in a blog post to have fixed the problem... gender-specific translations. The problem was… 1/3
This Post is from an account that no longer exists. Learn more
1
3
10
…all they really did was make it harder to illustrate the issue. “O bir doktor” would now give you a choice of male or female pronouns. But without any context either option is at least correct. If you provide context, it still gets it wrong and shows off its bias 2/3
1
1
2
This is a triple fail on google’s part: 1) outputting incorrect translations 2) failure to actually deal with the bias problem 3) masking the bias problem by cherry-picking and fixing well-known examples that can no longer be used to show the problem. 3/3
1
2