I have a more cynical view. GPT-5 is garbage and that’s why it hasn’t been released.
I base this on two things. First, in the hype-laden world of LLMs, no news is bad news. Second, we’ve seen this yarn spinning before with Google. Back in the early 2000s when Google could do no wrong, they were rumored to be in all kinds of sexy projects including a lot of cutting edge AI. Some turned out true, but most didn’t.
Later came the phase where their projects started dying away (there are many dozens of failed Google projects out there). And now, it’s to the sarcastic “do no evil” phase where their basic search service gets worse instead of better.. And we’re still waiting for that awesome AI stuff that Google was supposed to be into.
So when I hear of GPT-5 not being released on time, I don’t think “global takeover” or the “Singularity”. I think “another failed project”.
The only thing that is certain, the greater the money and power at stake, the greater the scheming and backstabbing. Using this filter, the claims made at the link have a high likelihood of being true.
LLMs so far have been a solution in search of a problem, and unfortunately for the technology, the solution doesn’t even solve the problems it’s been given, at least not without “hallucinations”.
Legislation is kicking around Congress to “ban” DeepSeek. The justification is that the AI would expose everything we use it for to the Chinese.
That would be a good argument, if it weren’t for the fact that you can download DeepSeek and run it on a system that is not only local, but not connected to the Internet at all. And it isn’t just a LLM: it’s a reasoning AI. Here’s a rundown by my favorite IT guy over at Dave’s Garage: bit.ly/3CPof0v
I have a more cynical view. GPT-5 is garbage and that’s why it hasn’t been released.
I base this on two things. First, in the hype-laden world of LLMs, no news is bad news. Second, we’ve seen this yarn spinning before with Google. Back in the early 2000s when Google could do no wrong, they were rumored to be in all kinds of sexy projects including a lot of cutting edge AI. Some turned out true, but most didn’t.
Later came the phase where their projects started dying away (there are many dozens of failed Google projects out there). And now, it’s to the sarcastic “do no evil” phase where their basic search service gets worse instead of better.. And we’re still waiting for that awesome AI stuff that Google was supposed to be into.
So when I hear of GPT-5 not being released on time, I don’t think “global takeover” or the “Singularity”. I think “another failed project”.
World Domination and Control! Why should an AI company be any different from any other company?
The same thing they’re after every evening, Pinky….
Floyd?
Narf!
The only thing that is certain, the greater the money and power at stake, the greater the scheming and backstabbing. Using this filter, the claims made at the link have a high likelihood of being true.
LLMs so far have been a solution in search of a problem, and unfortunately for the technology, the solution doesn’t even solve the problems it’s been given, at least not without “hallucinations”.
This may have something to do with why Musk wants to take control of Open AI.
Legislation is kicking around Congress to “ban” DeepSeek. The justification is that the AI would expose everything we use it for to the Chinese.
That would be a good argument, if it weren’t for the fact that you can download DeepSeek and run it on a system that is not only local, but not connected to the Internet at all. And it isn’t just a LLM: it’s a reasoning AI. Here’s a rundown by my favorite IT guy over at Dave’s Garage: bit.ly/3CPof0v