it's almost like we have a crystal ball or something, with @webdirections Code (Melbourne in June) covering passkeys, interaction to next pain and WebGPU, all key announcements at #GoogleIO2023 today
https://t.co/l6aijpSM5m
https://t.co/8ywBR20GLq
A demo I shared today in my keynote at @webdirections. It's a chat with Familiar that pulls from the web & combines this with a contextual memory that's been accrued locally about me over past interactions.
🗣️ Nota Bene to speakers: something I always do, is scan conference abstracts to see who I might be overlapping with. If it's the case, I try to reach out and see how we can compliment each other's work, *create cohesive content* + storylines. Had such a meeting today. Excited.
So stoked to be speaking in 2 days at @webdirections on `Breaking up long tasks`!
Wrote about it here: https://t.co/QVCPpHD3LH
Details here: https://t.co/Fcy7qMKJMH
Looking forward! @maxine@johnallsopp
This is one of these moments where current me is thanking past me: Not only do I keep drafts of potential talks, I also play w/ deck design ideas. Just grabbed one that's perfect, that I also sketched out almost 1yr ago. Thank you bored Henri! 🙃
I've long been more than sympathetic to the original Luddites ad recommend we all understand them a little better, more relevant now than ever.
A great place to start is Cory Doctorow's review of Brian Merchant's "Blood In the Machine
https://t.co/oJtHDJcyxZ
@simonw is one of the very best folks to follow, read and listen to when it comes to the impact of generative AI on software development, and this is an excellent podcast primer on LLMs
https://t.co/raeGuQv2qe
Since Twitter seem to have broken social media card previews (they only show the image and the domain now, not the title) here's how my TIL site's automated screenshot cards work, in case anyone wants to imitate that https://t.co/CP9sBSHwlC
Weapons-Grade
Guardrails, gaslighting and surfacing forbidden knowledge in Large Language Models
from @mpesce's new and highly recommended substack
https://t.co/P8718Inltw
@HenriHelvetica this is also yet another reason why you should never, ever, ever ever live code. The code is almost entirely illegible, plus what you think is an intelligible explanation is literally gibberish.
Trust me I edit lots of transcripts.