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

GPT-5

openai.com

Post reply on HN
Print view View original

941–950 of 1001 posts

Re: GPT-5

#941

Post by andsoitis »

somenameforme wrote: post #923
highfrequency wrote: post #331

It is frequently suggested that once one of the AI companies reaches an AGI threshold, they will take off ahead of the rest. It's interesting to note that at least so far, the trend has been the opposite: as time goes on and the models get better, the performance of the different company's gets clustered closer together. Right now GPT-5, Claude Opus, Grok 4, Gemini 2.5 Pro all seem quite good across the board (ie the…

If AGI is ever achieved, it would open the door to recursive self improvement that would presumably rapidly exceed human capability across any and all fields, including AI development. So the AI would be improving itself while simultaneously also making revolutionary breakthroughs in essentially all fields. And, for at least a while, it would also presumably be doing so at an exponentially increasing rate. But I thin…

> So the AI would be improving itself

Why would the AI want to improve itself? From whence would that self-motivation stem?

Re: GPT-5

#942

Post by cube00 »

accrual wrote: post #934

Already a lot of comments here (2188 at the time of this comment) but wanted to share my 2c: * It feels a bit more competent, as if it had more nuance or detail to say about each point. * It got a few obscure details about OpenBSD correct right away - both Sonnet 4 and 4o sometimes conflate Linux and OpenBSD commands. * It was fun asking GPT-5 to not only answer the query, but also to provide a brief analysis of the…

> It was fun asking GPT-5 [...] to provide [...] insights into myself!

Were those insights of a glowing and positive nature by chance?

Re: GPT-5

#943

Post by Argonaut998 »

They really nerfed Plus[0]. 80 messages every 3 hours for normal GPT-5. And only 200 messages per week for GPT-5 Thinking. It seems like terrible value.

Before it was: 100 o3 per week 100 o4-mini-high per day 300 o4-mini per day 50 4.5 per week

[0] https://help.openai.com/en/articles/11909943-gpt-5-in-chatgp...

Re: GPT-5

#944

Post by cj »

alexmorley wrote: post #910
ritzaco wrote: post #901

Ok this[0] sounds very, uh bold to me? Surely this is going to break a ton of workflows etc seemingly with nearly no notice? I'm assuming 'launches' equates with 'fully rolls out' or something but it's not that clear to me. When GPT-5 launches, several older models will be retired, including: - GPT-4o - GPT-4.1 - GPT-4.5 - GPT-4.1-mini - o4-mini - o4-mini-high - o3 - o3-pro If you open a conversation that used one of…

> For Free and Plus users, these changes take effect immediately. Pro, Team, and Enterprise users will also see the changes at launch but will have access to older models through legacy model settings. So only for free/plus users (for now). I do wonder how long they will take to deprecate these models via API though...

I have a ChatGPT Team plan, but the only model available is GPT-5. I'm not seeing an option to enable legacy models anywhere.

The only way to get access to other models right now (for me at least) is via the iPhone app, for now.

Re: GPT-5

#945

Post by nashashmi »

Argonaut998 wrote: post #943

They really nerfed Plus[0]. 80 messages every 3 hours for normal GPT-5. And only 200 messages per week for GPT-5 Thinking. It seems like terrible value. Before it was: 100 o3 per week 100 o4-mini-high per day 300 o4-mini per day 50 4.5 per week [0] https://help.openai.com/en/articles/11909943-gpt-5-in-chatgp...

Imagine the 90s being “you can only search x times a day”. Will we look back at the 20s in the same way.

Re: GPT-5

#946

Post by gooseus »

Neat, but it still fails my Forth simulator test, albeit a bit better than the last time. I hope they didn't have to raise the temperature of the ocean by too much to get these marginal improvements that definitely aren't even close to the kind of "general intelligence" you could get from training a clever child or monkey with 500 calories of sweets.

https://chatgpt.com/share/689525f4-20f0-8003-8bf6-f1f21dde6b...

You know what would be more impressive? If it said "Hey, I'm actually not designed to simulate a Forth machine accurately, I'm only going to be able to approximate it (poorly)). If you want an accurate Forth machine you should just implement this code: [Simple Forth Implementation]".

Or better yet, it could recognize when it was being asked to "be" a machine, and instead spin up a side process with the machine implementation and redirect any prompts to that process until a "STOP" token is reached.

Re: GPT-5

#947

Post by jstummbillig »

Argonaut998 wrote: post #943

They really nerfed Plus[0]. 80 messages every 3 hours for normal GPT-5. And only 200 messages per week for GPT-5 Thinking. It seems like terrible value. Before it was: 100 o3 per week 100 o4-mini-high per day 300 o4-mini per day 50 4.5 per week [0] https://help.openai.com/en/articles/11909943-gpt-5-in-chatgp...

If the past 2 years are any indicator cost per unit of capability will keep falling, rapidly.

Re: GPT-5

#948

Post by morelandjs »

Anyone know what the deal is with connectors? I don’t see them in the app and they made it sound like Google Calendar would be made broadly available as a connector.

Re: GPT-5

#950

Post by jltsiren »

somenameforme wrote: post #923
highfrequency wrote: post #331

It is frequently suggested that once one of the AI companies reaches an AGI threshold, they will take off ahead of the rest. It's interesting to note that at least so far, the trend has been the opposite: as time goes on and the models get better, the performance of the different company's gets clustered closer together. Right now GPT-5, Claude Opus, Grok 4, Gemini 2.5 Pro all seem quite good across the board (ie the…

If AGI is ever achieved, it would open the door to recursive self improvement that would presumably rapidly exceed human capability across any and all fields, including AI development. So the AI would be improving itself while simultaneously also making revolutionary breakthroughs in essentially all fields. And, for at least a while, it would also presumably be doing so at an exponentially increasing rate. But I thin…

There is no particular reason to assume that recursive self-improvement would be rapid.

All the technological revolutions so far have accounted for little more than a 1.5% sustained annual productivity growth. There are always some low-hanging fruit with new technology, but once they have been picked, the effort required for each incremental improvement tends to grow exponentially.

That's my default scenario with AGI as well. After AGI arrives, it will leave humans behind very slowly.

Post reply on HN