I was wrong about GPT-5
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Related videos
I’m done.
Theo - t3․gg
185.8k views
I wish this was clickbait
Theo - t3․gg
103.1k views
Prime is (mostly) right about AI
Theo - t3․gg
136.4k views
I don’t really like GPT-5.5…
Theo - t3․gg
132.8k views
I’m scared about the future of security
Theo - t3․gg
61.4k views
gpt-5.4 is really, really good
Theo - t3․gg
153.7k views
We need to talk about Sonnet 4.6
Theo - t3․gg
114.3k views
I was wrong
Theo - t3․gg
69.3k views
We need to talk about Ralph
Theo - t3․gg
109.9k views
I can't believe he was right.
Theo - t3․gg
64.4k views
Top Comments (10)
So the TL;DR is: He hyped GPT-5 because his early API build was 🔥, went on vacation during launch, came back to find the public version nerfed, everyone called him a shill, and now he’s defending his honor while OpenAI speedruns "How to Botch a Rollout."
Top 10 Anime Betrayals
Open Ai gives everyone a full taste of their latest models for 48 hours and then they nerf it and expect us not to notice.
Basically OpenAI turns up the power for reviews, and then detunes the engine because…. Shocker…. AI is not economically viable at the max level.
I told you guys Theo wasn't bribed, they nerfed the model.
I know this is WAY easier said than done, but I wouldn't take those comments too personally. Only you (and possibly close family/friends) will ever be able to know your true intentions and character. You've always struck me as a good guy, and I'd like to think I have a decent judgement of character... but I'm only exposed to ya through videos, like most others, so I can't say that with 100% certainty. But you seem like a good dude from what I've seen. Unfortunately, I think the types of reactions you're seeing stems more from where we are as a society than anything you've personally done. There are a lot of grifters and folks out there who are willing to set their morals and personal ideals aside in order to make a quick buck. So the kneejerk reaction from most folks is to treat everyone like they're grifters by default in order to "not be fooled", which is a shame, but also not surprising. A guard people have put up where they likely think it's better to not trust someone who is sincere than it is to have trusted someone who turned out not to be. Anyways... like I said, way easier said than done, but I'd try not to let it get to ya. If you've always given your genuine feelings about certain things, that's all you can do. Sure, maybe people can give advice on improving the timing, phrasing, or whatever... but what matters most is your intentions. If ya know in your heart that they were good, that's what matters most.
Unspoken rules when dealing with companies: 1. Do not believe anything a company promises until it is tangible 2. Don't forget the first rule
Respectfully, I think the issue here isn't/wasn't that you were authentic with your reactions or that you're a shill, but that you published what amounts to a product review of a pre-release version of a service without waiting to see if the actual final version matches what you'd seen and experienced. There's absolutely something to be said for wanting to be one of the first videos up to capitalize on the opportunity, but I think it's worth considering in future videos of the kind of thing that happened here with the preview version vs the reality and factoring that into future video scripts. I do think this is all very low stakes though and I understand your decision making 100% though and don't see that you have much blame aside from maybe needing to have couched your GPT-5 video with more "this is a preview the final version might be different, just look at what happened with llama" notes
30 min repeating the same 5 minutes of "poor Theo".
I think the best practice moving forward is to just not test things before they’re publicly released. Private demos always feature a controlled environment, and the only thing that matters is the final product.
Unlock the Data Inside
Turn Videos into Knowledge
- Get FREE 10/day: transcripts, summaries, chats
- Chat with videos, export text & PDF
- $1 free API credit for RAG, chatbots & research
Free forever plan • All features unlocked
Top Comments (10)
So the TL;DR is: He hyped GPT-5 because his early API build was 🔥, went on vacation during launch, came back to find the public version nerfed, everyone called him a shill, and now he’s defending his honor while OpenAI speedruns "How to Botch a Rollout."
Top 10 Anime Betrayals
Open Ai gives everyone a full taste of their latest models for 48 hours and then they nerf it and expect us not to notice.
Basically OpenAI turns up the power for reviews, and then detunes the engine because…. Shocker…. AI is not economically viable at the max level.
I told you guys Theo wasn't bribed, they nerfed the model.
I know this is WAY easier said than done, but I wouldn't take those comments too personally. Only you (and possibly close family/friends) will ever be able to know your true intentions and character. You've always struck me as a good guy, and I'd like to think I have a decent judgement of character... but I'm only exposed to ya through videos, like most others, so I can't say that with 100% certainty. But you seem like a good dude from what I've seen. Unfortunately, I think the types of reactions you're seeing stems more from where we are as a society than anything you've personally done. There are a lot of grifters and folks out there who are willing to set their morals and personal ideals aside in order to make a quick buck. So the kneejerk reaction from most folks is to treat everyone like they're grifters by default in order to "not be fooled", which is a shame, but also not surprising. A guard people have put up where they likely think it's better to not trust someone who is sincere than it is to have trusted someone who turned out not to be. Anyways... like I said, way easier said than done, but I'd try not to let it get to ya. If you've always given your genuine feelings about certain things, that's all you can do. Sure, maybe people can give advice on improving the timing, phrasing, or whatever... but what matters most is your intentions. If ya know in your heart that they were good, that's what matters most.
Unspoken rules when dealing with companies: 1. Do not believe anything a company promises until it is tangible 2. Don't forget the first rule
Respectfully, I think the issue here isn't/wasn't that you were authentic with your reactions or that you're a shill, but that you published what amounts to a product review of a pre-release version of a service without waiting to see if the actual final version matches what you'd seen and experienced. There's absolutely something to be said for wanting to be one of the first videos up to capitalize on the opportunity, but I think it's worth considering in future videos of the kind of thing that happened here with the preview version vs the reality and factoring that into future video scripts. I do think this is all very low stakes though and I understand your decision making 100% though and don't see that you have much blame aside from maybe needing to have couched your GPT-5 video with more "this is a preview the final version might be different, just look at what happened with llama" notes
30 min repeating the same 5 minutes of "poor Theo".
I think the best practice moving forward is to just not test things before they’re publicly released. Private demos always feature a controlled environment, and the only thing that matters is the final product.