Slide deck for my #monkigras talk 💬 "I Didn't Grow Up Speaking Code": GitHub Copilot as a Programming as a Second Language Tool 💬 is now live and accessible via my website!
Coming off #monkigras I can confirm it’s really popular for folks building AI products to say “AI should let humans be creative”, and I couldn’t agree more
but I’m wondering how often we’ve asked those same humans if we’re succeeding at that. Whether our solutions are giving them greater range of motion, or forcing them into a smaller box.
*for a very serious finding about how biased evaluation actually MASKS us from accessing achievement that might look the opposite to what we (stereotypically) expect. Aka, we need good math work done, and high achieving women could fill that demand, but our biases are directly sabotaging us from seeing the achievement of women in math skills.
A gift to get to share our science and evidence with this audience! We profoundly appreciate you welcoming our lab to the community #monkigras
To credit more craft, I am wearing my Mitchell Trousers which I made from a phenomenal pattern by Closet Core Patterns 😂 thus continuing the grand Dev Success Lab tradition where every time we present bespoke science we wear bespoke clothes we made
One of the delights of #monkigras is the tribe, and re-meeting old friends. Also, excited to learn that one of them is now working with one of my former coworkers! Lovely conversations, and cross-pollination of the community. Definitely an energising element of the event.
Current speaker is giving an example that gives me a deep sense of sadness. Using an AI agent to basically keep calling people to sell a product. Nope nope nope. This is where the usefulness of tools takes off in less palatable ways.
The fact that it is an Open Source thing doesn’t make me feel better about it.
I have very complex feelings about the state of the world at the moment.
These prospects are the most patient, friendly and sanitized responses to a cold call that I've ever seen. Makes me believe they were generated by a GenAI that is incapable of producing profanity, or even mild sarcasm. #monkigras
Interesting talk by Zack Akil highlighting some useful ways to use tools like Gemini. The “translate medical speak to everyday language” one was good, assuming it is validated by the medical professional. #monkigras
I know we can use AI to generate output and check for ourselves whether it’s good. But I do wonder what factors cause our own judgment to slip. Exhaustion? Overwhelm? The stress of deadlines?
And how do those things interact with there simply being MORE to review…thanks in part to gen AI? 🤔
You are applauding the AI that automates dealing with corporate BS now. But wait until the same AI will generate more corporate BS demands in six months from now #monkigras
Amazing talk (so far) by @grimalkina pointing out contest culture in different fields and the messaging around brilliance in those fields undermining women’s interest in them professionally #monkigras
One of themes I see in #monkigras talks so far is an exploration of gen AI and self-efficacy; where/when folks derive a sense of agency vs. fear of inadequacy from these tools.
We’re making ever-more sophisticated funhouse mirrors to ponder ourselves in.
Treating LLMs as a "friend who's read the Wikipedia article" ( @alex's phrase) is also helpful to find the right Wikipedia article in the first place. Yes, your friend is overconfident, and not as smart as they think, but they can still be helpful #monkigras