Hacker News

Sam Altman is in damage-control mode after latest ChatGPT release

Pffffft, this gonna be good.

OpenAI’s latest version of its vaunted ChatGPT bot was supposed to be “PhD-level” smart. It was supposed to be the next great leap forward for a company that investors have poured billions of dollars into.

Yeah, and suckers fell for the lie.

In short: It’s a dud.

It is not a dud. A dud is a device the creators expect to work; and it's clear "Open"A"I" never expected ChatGPT to be "PhD level smart".

1. It highlighted the many existing shortcomings of generative AI that critics were quick to seize on (more on that in a moment, because they were quite funny).

As in, it highlighted those critics were actually right? Go figure.

2. It raised serious doubts about OpenAI’s ability to build and market consumer products that human beings are willing to pay for. That should be particularly concerning for investors, given OpenAI, which has never turned a profit, is reportedly worth $500 billion.

Not a surprise here. The underlying technology is interesting, and you can find a few use cases for it; but it's certainly not profitable in the way those money-hungry and energy-hungry corporations think it should be. "I don't understand, we've been smearing the product on the customers' snouts all the time! And they still aren't swallowing it! Are those customers too stupid to understand simple concepts, like obedience? I wish we could force those things to do what we want, but sadly the junta we're backing up is not powerful enough to do so."

Now, one thing this industry is really good at is hype, and on that metric, CEO Sam Altman delivered.

Translation: industry is really good at being a bunch of filthy liars, and Sam Altman is a prime example of that.

During a livestream ahead of the launch last Thursday, Altman said talking to GPT-5 would be like talking to “a legitimate PhD-level expert in anything, any area you need.”

I trust Kika and Siegfrieda to reach that point before your product. Or you. (They're my cats, by the way.)

[Bluesky post] 500 billion dollars and the robot can't even count to twelve

At least it isn't trying to count the B's in «blueberry», or the R's in «strawberry».

But hey, once someone releases a pin-point patch to fix this stuff (much like whacking a mole), some muppet will say "I dun unrurrstand, it gets THIS ONE right! This must mean the system is smart!" - no, you muppet, it means you are not smart.

The bot returned an image of nine people instead, with rather creative spellings of America’s early leaders, like “Gearge Washingion” and “William Henry Harrtson.”

To be fair it seems to output actual letters now. Not random strokes. ...can someone test this shit with Cyrillic instead? I bet it'll get it even worse.

[Another Bluesky post] CEOkela

I don't even know which province that is supposed to be, but apparently CEOs are so important the local junta is naming provinces after them.

Especially because users have been particularly alarmed by the new version’s personality – or rather, lack thereof.

They waste a lot of time making the output sound like coming from a sycophant.

When people talk about AI, they’re talking about one of two things: the AI we have now — chatbots with limited, defined utility — and the AI that companies like Altman’s claim they can build — machines that can outsmart humans and tell us how to cure cancer, fix global warming, drive our cars and grow our crops, all while entertaining and delighting us along the way. // But the gap between the promise and the reality of AI only seems to widen with every new model.

The trick here is to realise one does not necessarily lead to the other. The "gap" there isn't just in time, but also functionality.

NN-based systems will stay, I guess. But corporations calling you stupid trash, by smearing lies on your face and expecting you to eat them up? I hope those are long gone from the field.