Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

1-800-ChatGPT

help.openai.com

Post reply on HN
Print view View original

221–230 of 350 posts

Re: 1-800-ChatGPT

#221

Post by conductr »

th0ma5 wrote: post #209
tracerbulletx wrote: post #181

Earlier quoted context omitted.

In every post about LLMs there is someone to blindly say something like this. When in reality if you ask ChatGPT for 10 good movies from this year you will get this. Anora - Directed by Sean Baker, a compelling drama about the life of a sex worker in Coney Island. Challengers - A provocative tennis drama directed by Luca Guadagnino, starring Zendaya. Dune: Part Two - Denis Villeneuve's continuation of the epic scienc…

Great! Now show me a system that can verify that list for accuracy as well. Not to be flippant, but this is the complaint. You can't approach outputs uncritically. And no I don't want it to be as unreliable as a person who also forgets how English or basic knowledge works at random intervals.

Movie reviews are amongst the most subjective things, how do you define “verify” and “accuracy”?

Re: 1-800-ChatGPT

#222

Post by JadeNB »

th0ma5 wrote: post #209
tracerbulletx wrote: post #181

Earlier quoted context omitted.

In every post about LLMs there is someone to blindly say something like this. When in reality if you ask ChatGPT for 10 good movies from this year you will get this. Anora - Directed by Sean Baker, a compelling drama about the life of a sex worker in Coney Island. Challengers - A provocative tennis drama directed by Luca Guadagnino, starring Zendaya. Dune: Part Two - Denis Villeneuve's continuation of the epic scienc…

Great! Now show me a system that can verify that list for accuracy as well. Not to be flippant, but this is the complaint. You can't approach outputs uncritically. And no I don't want it to be as unreliable as a person who also forgets how English or basic knowledge works at random intervals.

> Great! Now show me a system that can verify that list for accuracy as well. Not to be flippant, but this is the complaint. You can't approach outputs uncritically.

In general you can't, but surely it's not that big a deal if ChatGPT offers an inaccurate summary of a movie you're about to use to kill time on a flight? I suppose it becomes important if, e.g., you're relying on it to tell you whether a movie is appropriate for children, but, if you're just asking it whether a movie is worth watching, that's a question that doesn't have an objective, factual answer anyway, so a hallucinated answer is probably about as useful as that of a not-previously-known reviewer.

Re: 1-800-ChatGPT

#223

Post by glenstein »

tehwebguy wrote: post #204
tracerbulletx wrote: post #181

Earlier quoted context omitted.

In every post about LLMs there is someone to blindly say something like this. When in reality if you ask ChatGPT for 10 good movies from this year you will get this. Anora - Directed by Sean Baker, a compelling drama about the life of a sex worker in Coney Island. Challengers - A provocative tennis drama directed by Luca Guadagnino, starring Zendaya. Dune: Part Two - Denis Villeneuve's continuation of the epic scienc…

These details are already available in the in flight entertainment interface

They were responding to a commenter suggesting it would produce completely unusable results, the question was never about whether the results produced would be redundant.

I know that any mention of fallacies, valid or otherwise, causes instinctive eye rolls, but in this instance I agree with them that this amounts to moving the goalposts.

Re: 1-800-ChatGPT

#224

Post by elaus »

vletal wrote: post #177
solfox wrote: post #160

Calling it now: Google will ultimately lose this consumer battle. How? By doing what they've always done: building better tech. Gemini will be faster/better and it will have more features; but they will continue to fail to productize or explain it to consumers. Google's offerings here are still a huge mess. OpenAI is crushing them right now at building products that people want, and making them accessible.

Have you been paying attention at all these past few weeks? Google is crushing it with releases. Gemini 2.0 is great, Veo2 is crushing Sora, live video conversation from aistudio... 12 days of OpenAI turned out to be 12 days of Google.

> Google is crushing it with releases. Gemini 2.0 is great, […]

Isn't that exactly (part of) what they were saying in the comment you replied to?

> […] building better tech. Gemini will be faster/better and it will have more features

Re: 1-800-ChatGPT

#225

Post by beefnugs »

hansonkd wrote: post #147
jklinger410 wrote: post #68

Earlier quoted context omitted.

> but 60% of the time it completely lies to you This is such an exhausting conversation

i think when people say things like this it indicates that they tried LLMs in 2022 and solidified their opinion there. I had the same impression about the hallucinations 2 years ago. The reality is in at the end of 2024, you can get incredible value from LLMs. I've used copilot to code almost exclusively now for the past few months. Anyone still comparing it to text completion I feel is operating on completely out of…

Wait do you expect people to retry every failed thing they have tried due to marketing lies every how often exactly?

Re: 1-800-ChatGPT

#227

Post by conductr »

WhitneyLand wrote: post #77
cco wrote: post #5

100% shipped to show off to family during Christmas. I'll definitely give it a go, I wonder if this lands better with those aged 50+ who are more used to phone calls rather than chat.

So the perception of those aged 50+, is one of people so far removed from technology they’d prefer to use a telephone to avoid their discomfort with computers? I’m well into this group and still make a lot more api calls than phone calls. Fresh out of college I recall vividly thinking, I’ll need to build an impressive list of side projects to overcome preconceptions about how much I can truly offer at my age. Maybe n…

I think the SMS chat feature this enables is of more significance than the actual voice calling feature.

Re: 1-800-ChatGPT

#228

Post by sethhochberg »

vletal wrote: post #177
solfox wrote: post #160

Calling it now: Google will ultimately lose this consumer battle. How? By doing what they've always done: building better tech. Gemini will be faster/better and it will have more features; but they will continue to fail to productize or explain it to consumers. Google's offerings here are still a huge mess. OpenAI is crushing them right now at building products that people want, and making them accessible.

Have you been paying attention at all these past few weeks? Google is crushing it with releases. Gemini 2.0 is great, Veo2 is crushing Sora, live video conversation from aistudio... 12 days of OpenAI turned out to be 12 days of Google.

My respectful counterpoint is that most people aren't paying attention to tech releases at all, ever, unless they go viral like ChatGPT did.

I have very nontechnical coworkers get excited about cool new things ChatGPT can do, but I'm not certain any of them even know we _have_ Gemini in our Google Workspace.

This would hardly be the first time Google has produced innovative technology which eventually fizzles because it never captured much mindshare outside of the tech news circles

Re: 1-800-ChatGPT

#229

Post by stonedge »

rajamaka wrote: post #106
hypeatei wrote: post #49

Earlier quoted context omitted.

I agree about an abundance of apps, but what type of value are LLMs adding? It can sometimes be useful to input a more "human" search and have something get spit out but 60% of the time it completely lies to you. I'm talking about questions related to web specifications which are public documents. Section numbers, standards names, etc.. will be completely made up.

The value to me is by having an on-demand junior developer working alongside me for the price of $20 a month

I have upgraded to the $200 Pro tier, and, with o1-pro, all of my tasks delegated to the "junior" have been so much better. It takes longer to complete, of course, but the overall duration is less because I'm not having to go back and correct it as much as I was with 4o. It's been able to figure out problems that 4o continually failed on.

Re: 1-800-ChatGPT

#230

Post by freedomben »

Anyone know (and willing to share) what are they using for PSTN integration?

I've been looking into options for our non-profit tech startup (Ameelio) and about the best pricing I can find is about 1.35 cents per minute. It surprises (and saddens) me that it's still so expensive. I'm sure at a bigger scale you can negotiate better pricing, but based on the quick conversations I've had with vendors it doesn't get significantly cheaper.

Post reply on HN