I'm watching Nerd Snipe and hear that Gemini has a good load of world knowledge, like any Google model. But if it can't use that world knowledge, what would it be good at?
My first thought would be prompting other models: because it has that amount of world knowledge, it must know how to prompt models, right? Am I wrong here? What do you guys think?
I have access to iPhone mirroring now. It's pretty good, actually. This entire post was made from my iPhone, using my Mac and using iPhone mirroring. Pretty useless, but still fun to do.
I've seen people saying that GPT-6.1-Sol is not a good model, but it's near Astra performance for a much lower cost per task, and it's still the same token efficiency. So why do people say that GPT-6.1-Sol is not a good model?
Can I rant for a second? Why is there no split-screen feature on standard iPhones?
I would love to have something like that so I can chat with my friends and scroll on X at the same time. I hope this eventually gets implemented, but I don't know if @Apple will ever see this. Let's hope for the best, I guess.
This is my reaction to his Y Combinator post. I hope he sees this.
Also, yes, by the way: a face reveal. How stunning. But whatever.
Again, I don't mean this to hate on him, but I want people and him to view it as useful feedback.
I submitted @trySynara to @ycombinator
This is round 2, and the last shot I'm gonna give at YC
Whatever happens, I'm happy and proud of the work I've put in ❤️
So I forked T3 Code. I never realized how much fun it would be. It's so cool. I can change it to whatever I want!
Surely I'm not the first one to realize how much fun this is, but it's still cool that I'm able to do this with the help of AI.
Also, the UI isn't coherent at all across the app. You have way too many views for a lot of different things, and it feels a bit bolted on all over the place. I'm truly astonished.
Also, no cloud box support (e.g boat.dev)
Synara still needs a bit of polishing @emanueledpt
Hey Tibo, for your 28-day feature marathon or reset marathon, I'd love to see a new feature: a way to see how much I've spent on API costs if I hadn't had a ChatGPT subscription.
1/ I just discovered something that maybe everyone's already discovered. I was like to Codex, "Hey, my development process has been very slow. Can you use my computer history to figure out where the pain points are, why that is, and maybe work around that?"




