Woman on the tube has a phone in each hand.
Right hand has an email from her boss.
Let hand has ChatGPT.
She scrolls the e-mail into view, then has ChatGPT take a photo of it.
She gives the LLM a prompt, waits for the answer, and then thumb-types it into a reply.
I'm on my way to the #AIforTheRestOfUs conference. Didn't expect such an early demonstration!
The thing about AI conferences is that the BPP (Buzzwords Per Poster) is not demonstrably higher than any other conference.
I like bringing my own (eInk) namebadge to conferences.
But which QR code do you think is more dangerous to scan?
I'm in awe at Hannah Foxwell's ability to create diverse and inclusive events. Take so much work, but makes for a much better conference.
There are some good talks at #AIforTheRestOfUs but they're often undermined by crappy AI illustrations.
I've written about why people don't care about the quality of what they create or consume.
https://shkspr.mobi/blog/2024/12/most-people-dont-care-about-quality/
The problem with some of the "Vibe Coding" demos is that 90% of them could be replaced with just downloading a preexisting template.
The other 90% are caused by the ridiculous complexity of modern development stacks.
https://shkspr.mobi/blog/2020/06/weve-built-a-towering-pile-of-shite/
The Machine Learning stuff is excellent.
https://www.zoa.io/ is taking data from dozens of in-home sensors, across thousands of homes, with years of historic information - and creating a way to predict energy usage.
Beyond the LLM hypetrain, there's a quiet revolution.
@Edent How much energe did they use to collect the data and predict future energy usage?
@dpwiz do you understand how machine learning works, or are you confusing it with generative AI?
@dpwiz what a delightful interaction! Welcome to mute-town. Population: you.