I was more referring to joint and disc issues.
City where I live, MRIs are prohibitively expensive and take 6-8 weeks for availability…
Disruption can decrease prices, promote more choices, etc. I get your points of seeing too much/often and the negatives there…but hard for me to just say it’s not a good thing overall.
Trying new tools and building real things is still the fastest way to learn “AI”
It’s also how you develop and improve your “taste” too. No course needed. Just fumble around like an idiot and eventually the best practices become second nature.
Hahahaha. Google’s claims against SerpAI for scraping search results have been dismissed because…wait for it…Google can’t demonstrate it’s acting on behalf of the copyright holders.
In other words, if Google wants to refile the suit within the allowed 21 days, it has to admit that site owners have copyright protection of their work and THAT would open the door to them suing Google for scraping their content for AI Overviews.
Which reminds me: I’ve been meaning to reread Heller’s Catch-22. Now seems like a good time.
https://t.co/beZIfAeWQe
43 years ago Today, the movie WarGames was released.
A fascinating snapshot of Western popular culture in the 1980s, celebrating the home computing pioneering years
Going after dev audience. Faster, less taxing for both sides, different than the other cli offerings than the others.
V1 but with something new. Testing with their devout audience or highest subscribers. Makes sense and can iterate quickly. GUI comes faster after this is built as foundation.
I’d recommend using transcripts of the audio vs (only) audio. Whispr Flow, granola, whatever. Include both audio and written transcriptions.
When it comes to the copy, use Claude chat and set it to planning mode.
It’s always worth doing the same in ChatGPT.
Finalize copy and then you can import for use later on. Content and design shouldn’t be done at same time.
Open Claude Design to define design system on previous designs you’ve created or procured, ask it to ask you clarifying questions, spend 15 minutes getting it 80/20 there. Include style guide, fonts, colors. Make sure to give things you want to exclude.
Ask design to make the presentation in html. Often is part hrml and jsx. (Content is added as source to put into the presentation)
Can always take that and rework it backwards to a pdf or presentation.
After you make one in Claude design, it’s pretty plug and play to just rinse and repeat in my experience. Works much like the project feature in most tools.
There are ways to make it one shot and tools like Gamma are decent but easier to just spend sometime on the front end of the config. Results usually better and repeatable
You don’t have it backwards. They are going to focus on whales first. Capture that market for press.
Give their big name clients access to preview or beta version of models before general audience.
Then safe to assume that they will create a marketplace for the smaller providers. These companies will partner with OpenAI to start and share their services for social proof.
Likely same thing, being an smb partner will grant where premium access to models, especially if they exclusively use OpenAI and usage is at ## rate per month/clients etc
OpenAI will use this to pick out the best providers on the “small biz” side and they will acquire a meaningful chunk of those to own both sides.
Seems like an obvious path to me.
Today is a hard day. I shared this note with the @linear team today: We’ve made the difficult decision to increase our workforce. This is not a cost-cutting exercise or a reflection of anyone’s performance. We’re simply reimagining every role for the agentic AI era. We’re hiring. We’re sorry about that.