I've found it outstanding at isolated long running tasks (eg completed one of our tests in 3 hours and a 100% accuracy score versus 5.5 xhigh's 10 hours and 90% accuracy). For short tasks it seems very Claude'y (hard to express exactly what I mean by that) which I'm not a fan of meaning I'll stick with Codex for that use case and maybe Fable for those times I can for sure benefit from it.
Let's say you could follow and unfollow once a second and they knew it. Then you could replicate the dots and dashes by the length of time you toggled the status plus the gaps where you weren't toggling at all.
Eg in the first ten seconds 10 toggles, in the next 10 seconds a 5 second gap then 5 toggles. That would be dash, quiet, dot.
It's worth checking who the author is... Cristian's not exactly new to the game. I think he's being humble he doesn't know something despite his experience.
Thank you, that's very helpful. I think I could get close to that in some coding sessions where I'm running multiple in parallel but I suspect it's very very rare. Even with token efficient gpt5-codex my OpenAI bill is quite high so I think I will switch to Pro now.
I'm thinking about switching to ChatGPT Pro also. Any idea what maxes it out before I need to pay via the API instead? For context I'm using about 1b tokens a month so likely similar to you by the sounds of things.
Same. 30m tokens in a few hours (of which many were cached it seems and only a few output tokens). I use Gemini to solve only specific problems and in my case $20+ was worth it.
One of my favourite exchanges with Gemini saw it arguing with me that what I was asking of it was wrong. It culminated with it declaring (approximately) "I cannot in good faith continue to design for you something that is objectively flawed" and then refused to go further. Of course, Gemini was wrong and when I proved it without doubt it replied in a style that suggested it "felt" shame. Quite odd and very funny.
Recommendation: add more info to the home page to describe what it does without requiring sound. To evaluate this in a work environment I'd have to use my headphones which is an extra step I might not take unless the valid add and the product itself are very clear.
One of my first jobs (~1999) was as a research assistant working on seeing whether it's possible to make reentry vehicle heat shields from oak. Sadly, it wasn't.
I used magic patterns for a couple of months and it was one of the first no brainer AI services I've paid for outside of the main LLMs and IDEs. It did such an amazing job on quite an esoteric frontend that's very much not your normal web app. Impressive. Next time I need to design and build some more frontend code I'll be subscribing again.
Edit: to add some meat to that comment what surprised me was just how much better it was than Anthropic and OpenAI tools at that time for coming up with great looking products with minimal prompting. I also fed it other designs for inspiration and it replicated them brilliantly while incorporating my requirements. Good stuff.
I've found an air fryer on dehydrate works well. The Ninja I use controls temp within a few degrees and will happily do so for hours at up to 80C (actual recorded 80C unlike my filament dryer). Also relatively cheap and compact. Worth looking into. Only downside is there's not much space inside the drawer and you have to keep turning it to get an even drying as the air circulation is restricted.
This is great (in theory). During lockdown I got an ambisonic mic (Rode NF-SFW1) and used it to create Dolby Atmos experiences. The workflow - including sending it to Dolby's tool every time - was such a pain. Adding additional 3d elements was especially annoying and limiting.
Unfortunately that's no longer my hobby so can't test this for you but definitely scratches an itch for past me. Nice
R1 is the first model I've used that one-shotted a full JavaScript tetris with all the edge-case keyboard handling and scoring. It also one-shotted an AI snake game. With the right prompts I've found it consistently better than o1 and Claude 3.5 Sonnet.
The other thing is you'll need markers. At least in my experience. I think if you don't need precision photogrammetry and AI will get you the results way faster
(To be clear I love my Vega. Game changer. Just a bit heavy handed if you're not making car body parts)