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

ChatGPT Pro

openai.com

Post reply on HN
Print view View original

951–960 of 1001 posts

Re: ChatGPT Pro

#951

Post by carbocation »

katamari-damacy wrote: post #892
carbocation wrote: post #606

Earlier quoted context omitted.

Pro does have longer context windows , specifically 128k. Take a look at the pricing page for this information: https://openai.com/chatgpt/pricing/

ChatGPT and GPT4o APIs have 128K window as well. The 32K is from the days of GPT4.

According to the pricing page, 32K context is for Plus users and 128K context is for Pro users. Not disagreeing with you, just adding context for readers that while you are explaining that the 4o API has 128K window, the 4o ChatGPT agent appears to have varying context depending on account type.

Re: ChatGPT Pro

#952

Post by ToDougie »

__MatrixMan__ wrote: post #668
mtlynch wrote: post #664

Earlier quoted context omitted.

Kagi Ultimate (US$25/mo) includes unlimited use of all the Anthropic models. Full disclosure: I participated in Kagi's crowdfund, so I have some financial stake in the company, but I mainly participated because I'm an enthusiastic customer.

I'm uninformed about this, it may just be superstition, but my feeling while using Kagi in this way is that after using it for a few hours it gets a bit more forgetful. I come back the next day and it's smart again, for while. It's as if there's some kind of soft throttling going on in the background. I'm an enthusiastic customer nonetheless, but it is curious.

This is my biggest gripe with these LLMs. I primarily use Claude, and it exhibits the same described behavior. I'll find myself in a flow state and then somewhere around hour 3 it starts to pretend like it isn't capable of completing specific tasks that it had been performing for hours, days, weeks. For instance, I'm working on creating a few LLCs with their requisite social media handles and domain registrations. I _used_ to be able to ask Claude to check all US State LLC registrations, all major TLD domain registrations, and USPTO against particular terms and similar derivations. Then one day it just decided to stop doing this. And it tells me it can't search the web or whatever. Which is bullshit because I was verifying all of this data and ensuring it wasn't hallucinating - which it never was.

Re: ChatGPT Pro

#953

Post by everybodyknows »

lolive wrote: post #824
danpalmer wrote: post #515

Earlier quoted context omitted.

The problem is that OpenAI don't really have the enterprise market at all. Their APIs are closer in that many companies are using them to power features in other software, primarily Microsoft, but they're not the ones providing end user value to enterprises with APIs. As for ChatGPT, it's a consumer tool, not an enterprise tool. It's not really integrated into an enterprises' existing toolset, it's not integrated int…

Won’t name my company, but we rely on Palantir Foundry for our data lake. And the only thing everybody wants [including Palantir itself] is to deploy at scale AI capabilities tied properly to the rest of the toolset/datasets. The issues at the moment are a mix of IP on the data, insurance on the security of private clouds infrastructures, deals between Amazon and Microsoft/OpenAI for the proper integration of ChatGPT…

Is the Foundry business the reason for the run up of PLTR this year?

https://www.cnbc.com/quotes/PLTR

Re: ChatGPT Pro

#954

Post by freedomben »

TrackerFF wrote: post #390
tedsanders wrote: post #83

Earlier quoted context omitted.

No, o1 is definitely not good enough to replace employees. The reason we're launching o1 pro is that we have a small slice of power users who want max usage and max intelligence, and this is just a way to supply that option without making them resort to annoying workarounds like buying 10 accounts and rotating through their rate limits. Really it's just an option for those who'd want it; definitely not trying to push…

Good enough to replace very junior employees. But, then again, how companies going to get senior employees if the world stops producing juniors?

Indeed, I'm very concerned about this. Though i think it's a case of tragedy of the commons. Every company individually optimizes for themselves, fucking us over in the aggregate. But I think any executive arguing for this would have to be a pretty big company with an internal pipeline and promoting within to justify it, especially since everyone else will just poach your cultivated talent, and employees aren't loyal anymore (nor should they be, but that's a different discussion).

Re: ChatGPT Pro

#955

Post by jakubtomanik »

submeta wrote: post #394

I actually pay 166 Euros a month for Claude Teams. Five seats. And I only use one. For myself. Why do I pay so much? Because the normal paid version (20 USD a month) interrups the chats after a dozen questions and wants me to wait a few hours until I can use it again. But Teams plan gives me way more questions. But why do I pay that much? Because Claude in combination with the Projects feature, where I can upload two…

Have you tried LibreChat https://www.librechat.ai/ and just use it with your own API keys? You pay for what you use and can use and switch between all major model providers

Re: ChatGPT Pro

#956

Post by manmal »

apugoneappu wrote: post #895
danpalmer wrote: post #515

Earlier quoted context omitted.

The problem is that OpenAI don't really have the enterprise market at all. Their APIs are closer in that many companies are using them to power features in other software, primarily Microsoft, but they're not the ones providing end user value to enterprises with APIs. As for ChatGPT, it's a consumer tool, not an enterprise tool. It's not really integrated into an enterprises' existing toolset, it's not integrated int…

ChatGPT does have an enterprise version. I've seen the enterprise version with a top-5 consulting company, and it answers from their global knowledgebase, cites references, and doesn't train on their data.

Do you mean Azure OpenAI? That would be a Microsoft product.

Re: ChatGPT Pro

#957

Post by anticensor »

danpalmer wrote: post #547
ssl-3 wrote: post #332

Earlier quoted context omitted.

Does an o1 query run on a singular H100, or on a plurality of H100s?

A single H100 has 80GB of memory, meaning that at FP16 you could roughly fit a 40B parameter model on it, or at FP4 quantisation you could fit a 160B parameter model on it. We don't know (I don't think) what quantisation OpenAI use, or how many parameters o1 is, but most likely... ...they probably quantise a bit, but not loads, as they don't want to sacrifice performance. FP8 seems like a possible middle ground. o1 i…

o1 is ten times as expensive as pre-turbo GPT-4.

Re: ChatGPT Pro

#958

Post by unshavedyak »

torginus wrote: post #94
unshavedyak wrote: post #34

Earlier quoted context omitted.

I've not found value anywhere remotely close to this lol, but i'd buy it to experiment if they had a solid suite of tooling. Ie an LSP that offered real value, maybe a side-monitor assistant that helped me with the code in my IDE of choice, etc. At $200/m merely having a great AI (if it even is that) without insanely good tooling is pointless to me.

I don't know about you, but I get to solve algorithmic challenges relevant to my work approximately once per week to once per month. Most of my job consists of gluing together various pieces of tech that are mostly commodity. For the latter, Claude is great, but for the former, my usage pattern would be poorly served by something that costs $200 and I get to use it maybe a dozen times a month.

For me i feel like most of my time is spent inventing bespoke solutions in existing infra. Less about algorithms and more about making it work in an existing complex code base, which option will have the most negative impact, best impact, performant, etc.

A lot of tradeoffs to evaluate and it can be tiring onboarding people, let alone onboarding an AI.

Maybe it would massively improve my job if the AI could just grab the whole codebase, but we're not there yet. Too many LOC, too much legal BS, etc.

Re: ChatGPT Pro

#959

Post by lm28469 »

sebzim4500 wrote: post #348
lm28469 wrote: post #142

Earlier quoted context omitted.

> can kill any profit margins for the product as a whole. Especially when the base line profit margin is negative to begin with

Is there any evidence to suggest this is true? IIRC there was leaked information that OpenAI's revenue was significantly higher than their compute spending, but it wasn't broken down between API and subscriptions so maybe that's just due to people who subscribe and then use it a few times a month.

> Is there any evidence to suggest this is true?

I can't find any sources _not_ mentioning billions of loss for 2024 and for the foreseeable future

Re: ChatGPT Pro

#960

Post by __MatrixMan__ »

handfuloflight wrote: post #721
__MatrixMan__ wrote: post #668

Earlier quoted context omitted.

I'm uninformed about this, it may just be superstition, but my feeling while using Kagi in this way is that after using it for a few hours it gets a bit more forgetful. I come back the next day and it's smart again, for while. It's as if there's some kind of soft throttling going on in the background. I'm an enthusiastic customer nonetheless, but it is curious.

Is that within the same chat?

The flow lately has been transforming test cases to accommodate interface changes, so I'm not asking it to remember something from several hours ago, I'm just asking it to make the "same" transformation from the previous prompt, except now to a different input.

It struggles with cases that exceed 1000 lines or so. Not that it loses track entirely at that size, it just starts making dumb mistakes.

Then after about 2 or 3 hours, the size at which it starts to struggle drops to maybe 500. A new chat doesn't seem to help, but who can say, it's a difficult thing to quantify. After 12 hours, both me and the AI are feeling fresh again. Or maybe it's just me, idk.

And if you're about to suggest that the real problem here is that there's so much tedious filler in these test cases that even an AI gets bored with them... Yes, yes it is.

Post reply on HN