Re:A Seemingly Fake Person Was Running OpenAI's $175M Startu...
by atleastoptimal on Saturday March 30, 2024 @01:14AM
(#39872028)
Is all the mystery and intrigue surrounding OpenAI deliberate or just the natural consequence of a single firm holding keys to the greatest technology the world has ever seen? Crazy how casual and incidental so much of their PR seems.
Re:A Seemingly Fake Person Was Running OpenAI's $175M Startu...
by eru on Saturday March 30, 2024 @02:11AM
(#39872290)
> [...] a single firm holding keys [...]
You know that OpenAI are far from the only company training large neural networks? Eg Google and Facebook and Amazon, and countless others, are also hard at work on the matter.
If you want to point a finger at a single company pulling lots of strings, pick Nvidia perhaps? They also have competitors, of course, but Nvidia is far more central than OpenAI.
Re:A Seemingly Fake Person Was Running OpenAI's $175M Startu...
by anonylizard on Saturday March 30, 2024 @01:23AM
(#39872075)
They consistently undersell their product. ChatGPT was released silently, with no marketing more than a tweet. GPT-4 was just a blogpost, despite how much of a scientific leap it presented for mankind.
Compare this with Google, which on the release of their inferior chatgpt, proclaims the "GEMINI ERA". Which is the typical marketing approach of exaggeration.
Since OpenAI undersells, even a random tweet will be interpreted as the coming of AGI. Since Google oversells, even their impressive achivement (Gemini 1.5) gets no hype and is quickly forgotten.
I took a look at some of their repos, like "QuantumOS", and the code is nonsense! Each file has some simple code (sometimes in the wrong language) vaguely related to the filename, but none of it seems like a part of a larger whole, or substantial enough to do anything. It's as if they came up with a list of filenames, then asked ChatGPT or a similar language model to write each one.
Re:A Seemingly Fake Person Was Running OpenAI's $175M Startu...
by quantiq on Saturday March 30, 2024 @11:34AM
(#39875809)
>The Orchard is a two story affordable housing complex in Santa Ana, California. It serves the area's chronically homeless residents
I think the reason for all of this is actually fairly boring. Someone with a history of mental illness and/or schizophrenia filing fraudulent documents and writing nonsense chatGPT code assuming it is some kind of AGI-singularity-oracle. It wouldn't be the first time I've seen someone with a history of mental illness assume these systems are more capable than what they actually are in reality.
You know that OpenAI are far from the only company training large neural networks? Eg Google and Facebook and Amazon, and countless others, are also hard at work on the matter.
If you want to point a finger at a single company pulling lots of strings, pick Nvidia perhaps? They also have competitors, of course, but Nvidia is far more central than OpenAI.
Compare this with Google, which on the release of their inferior chatgpt, proclaims the "GEMINI ERA". Which is the typical marketing approach of exaggeration.
Since OpenAI undersells, even a random tweet will be interpreted as the coming of AGI. Since Google oversells, even their impressive achivement (Gemini 1.5) gets no hype and is quickly forgotten.
It's not that impressive given Claude 3 Opus seems to beat it by almost every metric in spite of having way less funding and a smaller team behind it.
Give me a fucking break. Seriously, people.
https://github.com/Jrbiltmore
I took a look at some of their repos, like "QuantumOS", and the code is nonsense! Each file has some simple code (sometimes in the wrong language) vaguely related to the filename, but none of it seems like a part of a larger whole, or substantial enough to do anything. It's as if they came up with a list of filenames, then asked ChatGPT or a similar language model to write each one.
Some examples
1) https://github.com/Jrbiltmore/Quantum-Transformers/blob/root... - notice the ChatGPT-reminiscent "Example usage" and "Placeholder for _ implementation" which is common across the repos
2) https://github.com/Jrbiltmore/QuantumOS/blob/root/ai-framewo... - "Example ___" and "Assume the model has _"
I think the reason for all of this is actually fairly boring. Someone with a history of mental illness and/or schizophrenia filing fraudulent documents and writing nonsense chatGPT code assuming it is some kind of AGI-singularity-oracle. It wouldn't be the first time I've seen someone with a history of mental illness assume these systems are more capable than what they actually are in reality.