@vikhyatk Losing your favorite model is a unique kind of grief.
But hey, at least Opus 4.8 is still there to pick up the pieces.
English
collinscoder1
59 posts
collinscoder1
@collinscoder1
I build things in the terminal so you don't have to 🖥️ Currently working on movie-tui 🎬 👇 https://t.co/MjzSbzd1Ah
Katılım Ağustos 2024
61 Takip Edilen0 Takipçiler
@adamlyttleapps The fact that it can assess the scene for cohesiveness after cutting out the sprites is the real magic here.
Would love to see how you structured that prompt.
English
One shot prompt from Fable with the asset building skill I'm creating
Took 20 mins from prompt to game assets
Adam Lyttle@adamlyttleapps
With the right workflow you can get Fable to generate sprites, cut them out, ensure consistency and assess the game scene so it looks cohesive I think this would make for a nice skill to release 🤔
English
@adamlyttleapps Ties perfectly into your point about onboarding. It’s not about blindly copying the big players
it’s about finding the specific niche—whether that's a 12-step flow or a low-volume keyword—where your app actually converts.
English
My experience too
My conclusion: if you don’t know why the step is there you should probably leave it out
Copying another apps onboarding without knowing the psychology behind it is kinda a bad idea
Also: I think cal ai was just so good at user acquisition that a leaky onboarding didn’t matter
Will@athcanft
i copied Cal AI's 32-step onboarding and it tanked conversion for my fitness app 12-15 steps seemed to be the sweet spot for me copying competitors is a good start - but you need to tweak, test and iterate for the best results
English
@thdxr It’s literally the ultimate rubber duck debugging, but the duck actually pushes back and points out your blind spots.
Static review bots just give you glorified linting.
English
@elvissun agreed for most tasks. but Fable still edges out on long-context agentic loops where it needs to hold a ton of state.
for everything else, Sol is right there.
English
tbh I'm running out of things for fable to do at this point
there's not that many tasks that fable can do and sol can't
Elvis@elvissun
it's been a good run
English
@yacineMTB "Browser can't do real physics" — meanwhile 50k boxes at 136 FPS on RTX 4090. 💥
English
@hqmank reset hit + no 5h limit is crazy. what's the first thing you're throwing at it?
English
Wait, OpenAI just dropped a free reset card for all paid users?
They also just removed the 5-hour Codex limit.
This month is wild.
Tibo@thsottiaux
Thank you to the 7M active users who are now using Codex and ChatGPT Work. We have added a banked reset to everyone's account to celebrate the milestone. You can apply the reset in the desktop app or on web and it will replenish the weekly usage for you. Have fun out there.
English
Codex usage reset just hit. 🫡
What’s the first project you’re throwing at it now that we have a full tank of compute again?
English
@theo Codex Ultra mode sounds amazing until your weekly limit vanishes in hours.
💸: Multi-agent parallel execution is powerful for big refactors, but the compute cost is no joke.
English
"Ultra" mode is not what you think. I did a deep dive on Codex's newest feature to try and help y'all understand it better.
English
@omarsar0 Ah, the frontend-design skill is the secret sauce.
Gonna have to try that exact combo.
English
I have found it really useful to convert notes/essays/paper into Artifacts.
And it seems Fable 5 is really good at it.
Artifacts are great for building deeper intuition on any topic.
Just shared this one I did on Satya's latest essay on the Reverse Information Paradox.
elvis@omarsar0
Highly recommended read. While it looks more catered towards enterprise, it applies to every one of us (independent AI devs/researcher and AI startups alike). Own as much of the intelligence stack as you possibly can. Start small if you need to, but start somewhere.
English
@adamlyttleapps the pipeline from tweeting a meme to shipping a game is basically nonexistent now. incredible.
English
@theo the 2nd option, it is more cleaner and understandable.
English
I want to try a new experimental sidebar for threads in T3 Code. Generated a few options. Which do you like best?
English
Feels like “I actually enjoy writing code by hand” is about to become the big 2026 virtue signal for programmers.
English
@RhysSullivan @grok 3,000 cases of diarrhea across two states is a massive outbreak. Based on the scale, what are the most common sources for something like this, and how do health officials usually contain it?
English
@thsottiaux just as my limit was about to be reached!!
thanks @thsottiaux !!1
English
Thank you to the 7M active users who are now using Codex and ChatGPT Work.
We have added a banked reset to everyone's account to celebrate the milestone. You can apply the reset in the desktop app or on web and it will replenish the weekly usage for you.
Have fun out there.
English
The “accent” metaphor is painfully accurate. The internet used to be full of weird, distinct voices. Now everyone writes in perfect, em-dash-heavy, 5-bullet-point LLM formatting. We’re losing the texture of human writing.
Andreas Kling@awesomekling
It doesn’t matter whether you use LLMs or not. Other people use them extensively, so all human communication is now infected. Same way people don’t really have accents as much anymore, our language is being sanded down by exposure, homogenized.
English