Felipe Rezende

[Felipe Rezende @FelipeBSR]

@bindureddy Creative writing?

[Felipe Rezende @FelipeBSR]

@adrian_valentim Dei sorte, peguei uma segunda conta, deixei uma dezena de prompts engatilhados e consegui passar uma semana usando. Meu uso é diferente do seu, então não tive muitas restrições, achei a capacidade dele excelente. Se voltar a ficar disponível, deverá ter ainda mais guardrails.

[Felipe Rezende @FelipeBSR]

@eixopolitico Muito bizarro, muita gente migrou de plano para usar mais o Fable e agora vem essa pedrada.

[Felipe Rezende @FelipeBSR]

@AnthropicAI WTF?

[Felipe Rezende @FelipeBSR]

@OpenAINewsroom @ona_hq And when are you going to release the 5.6?

[Felipe Rezende @FelipeBSR]

@HarshithLucky3 The day didn't end yet... but yea, it seems to not be today, no one leaked anything.

[Felipe Rezende @FelipeBSR]

@adamhfry Neat, but how is it going to affect the usage limits? Can you send the same number of messages on High vs on Medium? For the Plus Plan.

[Felipe Rezende @FelipeBSR]

@ChatGPTapp When will you release the 5.6?

[Felipe Rezende @FelipeBSR]

@haider1 The seer has spoken, so we're going to have GPT 6.

[Felipe Rezende @FelipeBSR]

@thsottiaux @thsottiaux can you pick 6 different numbers from 1 to 60? If i win the lottery will give you a generous tip, great seer.

[Felipe Rezende @FelipeBSR]

@sakurayukiai @cohere Local VRAM is the RAM from GPU or can you use from other sources like normal RAM?

[Felipe Rezende @FelipeBSR]

@minchoi And creative writing? Can you adjust its system prompt and make ir more flexible? Anyways, seems they're fusing codex to GPT in next update. I really hope easier use of agent mode for creative writing.

[Felipe Rezende @FelipeBSR]

@HarshithLucky3 Use 5.3codex. Oh wait, they removed it! It was a shame what they did.

[Felipe Rezende @FelipeBSR]

@koltregaskes @Pawankalhans Me too. Even if somehow it would be available to the $20 plans, the token drainage would be too high. Like one or 2 messages per 5 hours. Just a taste to make you go for the $200 plan. GPT is the pack mule and i hope it gets a boost too... excited about increases in agent mode.

[Felipe Rezende @FelipeBSR]

@VraserX - bigger context input size
- agent mode without too much limitations, maybe drawing from Codex's quota

[Felipe Rezende @FelipeBSR]

@NVIDIAGeForce #RTXPowersPlay would be perfect right now!

[Felipe Rezende @FelipeBSR]

@restlessmindxd @mindofjota O problema que vai custar um rim em tokens o mythos.

[Felipe Rezende @FelipeBSR]

@thetreygoff I'd say depends on the task and depends on Opus' version. At his moment, big drawback is the limit. Using both is great, copying output from one and placing in other's input. GPT is the pack mule and Opus the race horse. At least outsite codex/claude code axis.

[Felipe Rezende @FelipeBSR]

@NVIDIAGeForce #RTXPowersPlay

[Felipe Rezende @FelipeBSR]

@AnthropicAI More likely you're getting out of compute for stronger models and would love to OpenAI to stop progressing. That's why i love competition. If there was no OpenAI, Claude probably would be Sonnet only or up to Opus 3. And same thing with chatGPT. Keep improving and competing.