GPT-4 has a usage cap of 40 messages every 3 hours for Pro users. Once you run out, it'll put a message up that you're running on GPT-4o until the next time period.
@thatroblennon
-
GPT-4o Model Selection and Message Credit Management Issue
By
–
Really wish you could decide whether your GPT was using GPT-4o or not. Currently the only way to run 'em on 4o is to burn through all your higher value messages first. @OpenAI
-
Using ChatGPT and Poe for Business API Development
By
–
Both. API for my business (e.g. Reactor) but I primarily use ChatGPT and Poe .com
-
Using Sonnet for routine tasks to preserve Opus quota
By
–
No, generally it isn’t (although I’ve had some weird outliers where it outperforms Opus) I just prompt so much that I’ll use Sonnet for less important stuff so that I don’t max out my Opus usage
-
GPT-4o for Analysis, Claude Sonnet for Creative Tasks
By
–
My list is the basically the same although I use GPT-4o for my 'analytical' workhorse and Claude Sonnet for stuff that's more creative https://t.co/tBgnIJ4DXP
— Rob Lennon 🗯 | AI Whisperer (@thatroblennon) 26 mai 2024My list is the basically the same although I use GPT-4o for my 'analytical' workhorse and Claude Sonnet for stuff that's more creative
-
Black Magic Motivators in Prompts Without Right Answers
By
–
I've seen it work enough times now that I always include one "black magic" motivator, even though I don't always test to make sure it's actually helping. For a lot of my prompts, there is no right or wrong answer (such as business strategy or content creation), so it's
-
AI Gets Green Tea and Sleep Instead of Adderall Today
By
–
I'm sure you all will be relieved to know: Instead of Adderall I'm giving my AI green tea and a good night's sleep today
-
AI Businesses Available for Sale at Profitable Terms
By
–
If anyone would like to buy any of my AI businesses, I will also sell them for a similar price to profit ratio (PS – people actually like Reactor)
-
Experimenting with Larger Prompts on Cheaper Models
By
–
Now we're talkin. 🙂 I have quite a few 1,000-4,000 token prompts currently in use. But aiming to experiment with running cheaper models with bigger prompts that essentially contain a ton of examples. Fine tune without the fine tune.
-
Real Prompt Engineering: Tedious but Satisfying Problem-Solving
By
–
Real prompt engineering is tedious but satisfying. I just spent 6 hours experimenting to solve a handful of bugs in the Reactor .is ghostwriting pipeline. Sometimes you gotta try a dozen different techniques to find a way to solve an issue consistently and without creating new