Inside OpenAI's Turbulent Year
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
YouTube is Already 20% AI Slop
ColdFusion
146.7k views
So... What Was the Point of Artemis II?
ColdFusion
56.9k views
OpenAI Kills Sora then Descends into Chaos
ColdFusion
51.1k views
How The Internet’s Favourite AI Employee Went Rogue
ColdFusion
41.6k views
Yikes.
ColdFusion
350.1k views
OpenAI is Suddenly in Trouble
ColdFusion
665.5k views
OpenAI "Full Code Automation" Coming This Year...
Wes Roth
88.0k views
The Rise of Vietnam
ColdFusion
912.9k views
Australia's Insane Social Media Ban
ColdFusion
634.4k views
Samsung is in Crisis
ColdFusion
3.5m views
Top Comments (10)
The sacrifices whistleblowers make to expose inhumane corporations..
why whisterblowers die?
Zuck and Elon vs OpenAI? All I can say is that there will be no victory for the average person there.
The only prediction I have about AI is that a handful of individuals will get insanely rich.
The worst part is that this happens so often nowadays that you can never possibly tell if it was an accident or not. The fact that it happens so often... makes you scratch your head.
The frequency we talk about ‘leaks’ and ‘whistleblowers’ with a company named ‘Open’AI is 2024.
Whistleblowers are the ones who kept humanity alive ❤😢
We all need to give Dagogo more praise for this amazing channel. Thanks for staying true to the vision.
Something to consider that seems to have been omitted about the results shown at 8:35, is that the OpenAI model was allowed 50 submissions per question, while each of the human participants were only allowed a single final submission per question. This was also the case in the Math trials as well, where the OpenAI models were allowed hundreds or even thousands of attempts to solve the given problems. So while it is still very impressive that these machines are able to solve these problems in a comparatively short timeline when compared to a human, these are still well-structured problems with distinctly defined solutions. While "in the wild", and especially with non-trivial professional use-cases, AI output frequently requires heavy oversight from a field expert to analyze and verify its validity. Anyone whose toyed around with text, image, song, and code generation knows that it can require multiple attempts to get the desired output (or in some cases, even anywhere in the ballpark of desired output), and the more complex and technical the task, the higher the level of required expertise and scrutiny is needed to make sure the output is correct (or can be modified to be correct).
Whistleblowers don’t commit suicide. When you understand the mental fortitude necessary to face giants by whistleblowing, you understand why every whistleblower’s “suicide” is most likely murder.
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)
The sacrifices whistleblowers make to expose inhumane corporations..
why whisterblowers die?
Zuck and Elon vs OpenAI? All I can say is that there will be no victory for the average person there.
The only prediction I have about AI is that a handful of individuals will get insanely rich.
The worst part is that this happens so often nowadays that you can never possibly tell if it was an accident or not. The fact that it happens so often... makes you scratch your head.
The frequency we talk about ‘leaks’ and ‘whistleblowers’ with a company named ‘Open’AI is 2024.
Whistleblowers are the ones who kept humanity alive ❤😢
We all need to give Dagogo more praise for this amazing channel. Thanks for staying true to the vision.
Something to consider that seems to have been omitted about the results shown at 8:35, is that the OpenAI model was allowed 50 submissions per question, while each of the human participants were only allowed a single final submission per question. This was also the case in the Math trials as well, where the OpenAI models were allowed hundreds or even thousands of attempts to solve the given problems. So while it is still very impressive that these machines are able to solve these problems in a comparatively short timeline when compared to a human, these are still well-structured problems with distinctly defined solutions. While "in the wild", and especially with non-trivial professional use-cases, AI output frequently requires heavy oversight from a field expert to analyze and verify its validity. Anyone whose toyed around with text, image, song, and code generation knows that it can require multiple attempts to get the desired output (or in some cases, even anywhere in the ballpark of desired output), and the more complex and technical the task, the higher the level of required expertise and scrutiny is needed to make sure the output is correct (or can be modified to be correct).
Whistleblowers don’t commit suicide. When you understand the mental fortitude necessary to face giants by whistleblowing, you understand why every whistleblower’s “suicide” is most likely murder.