James Brady

The "Of One" Concept | Anyone can be an Army | Anyone can become a force to change the world | AI is raw creation for those that want to do, build, be ANYTHING

[James Brady @of1ai]

Dear OpenAI,

It's not okay that I have to use Opus 5.5 to get your configuration to work so that I can properly use Ultrafast. @OpenAIDevs @thsottiaux

I have paid for 3 500 accounts and each one is having issues for me. I switch to ultrafast on 6.1 and it says it is on it but then I have opus look at the tps, and the tps is not ultrafast... not at all.

Opus is currently working on another fix so I can use it again...

[James Brady @of1ai]

Bing now shows you the AI-answer queries where other sites get cited and yours doesn't. It's in Webmaster Tools under Topics Portfolio (beta).

Use it as a gap list, not a strategy. Strike the competitor brands, then ship one page per real gap.

https://t.co/pCPVvJw0EQ

[James Brady @of1ai]

@OpenAIDevs @thsottiaux @embirico @romainhuet 6/ Asks:
1. Serve Ultrafast to every account whose plan includes it.
2. Show which speed actually served each turn. Never downgrade silently.
3. Tell us if these turns were billed at Ultrafast rates, and credit them if so.

Paying for Ultrafast and getting Standard isn't OK.

[James Brady @of1ai]

@OpenAIDevs You have an on-going problem. I will post on it again. When I select Ultrafast (the flag indicates ultrafast), but the model being fed is on standard speed. I will post details soon. This has been an on going problem.

[James Brady @of1ai]

@OpenAIDevs @thsottiaux @embirico @romainhuet 5/ ChatGPT for Mac 26.1007.71326, Codex engine 0.162.0-alpha.17.2, macOS 15.7.4.

Account B also used 11% of its weekly limit in two hours this morning. We can't see whether those turns were billed at Ultrafast rates.

[James Brady @of1ai]

@natebjones Only 8k on here, and you're killing it on TikTok, huh? This is the first time you popped into my feed, and of course I added you.

A solid, even kilter source is not the ideal on X, I guess...

[James Brady @of1ai]

@TokenGremlin I like your content.

[James Brady @of1ai]

I use Claude, honestly, for my orchestration work. I would use it for everything, but it doesn't work with passwords, so I switch back to GPT. I use GPT for some automations, but I honestly don't trust 6.1 as much as I trust Opus to correctly and honestly articulate its work.

I'm a huge OpenAI fan, but honestly, with the usage difference, the limits, and everything, Anthropic is killing it right now. I have double the OpenAI accounts I have with Anthropic, and I am having trouble getting through my Anthropic limits. I can nuke GPT easily, and it fricking sucks because I'm a huge GPT fan. It's like, that's my team, bro. That's my team. I know we shouldn't have teams, but that's my team, and they are failing me.

[James Brady @of1ai]

@anah_sahh I have two start ups that are net positive six figures both started with AI. When is close to seven figures and I think it will scale to eight over the next 2 to 3 years

[James Brady @of1ai]

@rezoundous It’s true I have a ton of floating resets cause gpt is just not good enough to matter right now

[James Brady @of1ai]

@BratDotAI Model>opus 5.5

[James Brady @of1ai]

@anah_sahh I have met one. Same guy told me he was planing to build a company platform primarily by himself.

[James Brady @of1ai]

GPT 6.1 is durpy lately. Many key mistakes. Like dumb non-sense I would expect from opus 4.8

Completely overlooking requests. Not just dumb, just odd off…. Weird….

A client asked it to change a number on the website to a specific number and then it changed the number on the website to a different number and it’s done this multiple times across different client websites they’ve had to go back in with opus 5.5

[James Brady @of1ai]

@haider1 I really don't think that qualifies as a reason for their massive user base and real-world use. It's a rather ironic comment, considering Google's user base is so large because Google is synonymous with looking up something on the internet. That alone says everything.

[James Brady @of1ai]

@SamVale29 I would not hire a strict dev at all

[James Brady @of1ai]

@thsottiaux Dude, this is me literally every few days and I have Max Studios with 8 TB and I have a mixed studio with 4 TB. This is a constant issue.

I thought it was just me