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

981–990 of 1001 posts

Re: ChatGPT Pro

#981

Post by wkat4242 »

I really wish they would drop the Google play requirement on Android. I have Google play installed on my Samsung but it's just not logged in. You don't need to be logged in for a lot of functionality like push notifications.

Every single app works fine that way, except ChatGPT. It opens the play store login page then exits. I have no problems with apps from 2 banks, authenticators etc etc.

It's just so weird that they force me to make an account with one of their biggest competitors in AI. I just don't want to, I don't trust Google with my data. By not logging in they have some but not a lot.

iOS isn't an option either because it's too locked down. I need things like sideloading and full NFC access for things like OpenPGP.

Re: ChatGPT Pro

#982

Post by wkat4242 »

devjab wrote: post #572
fudged71 wrote: post #443

OpenAI is racing against two clocks: the commoditization clock (how quickly open-source alternatives catch up) and the monetization clock (their need to generate substantial revenue to justify their valuation). The ultimate success of this strategy depends on what we might call the enterprise AI adoption curve - whether large organizations will prioritize the kind of integrated, reliable, and "safe" AI solutions Open…

I wonder if OpenAI can break into enterprise. I don’t see much of a path for them, at least here in the EU. Even if they do manage to build some sort of trust as far as data safety goes, and I’m not sure they’ll have much more luck with that than Facebook had trying to sell that corporate thing they did (still do?). But if they did, they will still be facing the very real issue of having to compete with Microsoft. I…

> I wonder if OpenAI can break into enterprise. I don’t see much of a path for them, at least here in the EU.

Uhh they're already here. Under the name CoPilot which is really just ChatGPT under the hood.

Microsoft launders the missing trust in OpenAI :)

But why do you think copilot is worse? It's really just the same engine (gpt-4o right now) with some RAG grounding based on your SharePoint documents. Speaking about copilot for M365 here.

I don't think it's a great service yet, it's still very early and flawed. But so is ChatGPT.

Re: ChatGPT Pro

#983

Post by wkat4242 »

solarkraft wrote: post #609
CobrastanJorji wrote: post #579

Earlier quoted context omitted.

Man, if they can solve that "trust" problem, OpenAI could really have an big advantage. Imagine if they were nonprofit, open source, documented all of the data that their training was being done with, or published all of their boardroom documents. That'd be a real distinguishing advantage. Somebody should start an organization like that.

It's sort of funny how close they were to that until Altman came along.

Well and also the Microsoft billions. They had a lot to do with that as well. Once you're taking that kind of money you can't really go back.

Re: ChatGPT Pro

#984

Post by wkat4242 »

bradfox2 wrote: post #791
CobrastanJorji wrote: post #579

Earlier quoted context omitted.

Man, if they can solve that "trust" problem, OpenAI could really have an big advantage. Imagine if they were nonprofit, open source, documented all of the data that their training was being done with, or published all of their boardroom documents. That'd be a real distinguishing advantage. Somebody should start an organization like that.

The cyber security gatekeepers care very little about that kind of stuff. They care only about what does not get them in trouble, and AI in many enterprises is still viewed as a cyber threat.

One of the things that i find remarkable in my work is that they block ChatGPT because they're afraid of data leaking. But Google translate has been promoted for years and we don't really do business with Google. Were a Microsoft shop. Kinda double standards.

Re: ChatGPT Pro

#985

Post by wkat4242 »

ToDougie wrote: post #952
__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.

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…

Could it be that you're running out of available context in the thread you're in?

Re: ChatGPT Pro

#986

Post by wkat4242 »

ezst wrote: post #906
epigramx wrote: post #860

Earlier quoted context omitted.

And the catch-up of logic clock, how fast people catch-up we don't have Skynet within 2 years, but a glorified google search for the next 20 years.

Am I the only one who's getting annoyed of seeing LLMs be marketed as competent search engines? That's not what they've been designed for, and they have been repeatedly bad at that.

Yeah they're totally not designed for that. I'm also surprised that companies that surely know better market it as such.

Combined with a search engine and AI summarisation, sure. That works well. But batebones no. You can never be sure whether it's hallucinating or not.

Re: ChatGPT Pro

#987

Post by wkat4242 »

j45 wrote: post #707
johnisgood wrote: post #596

Earlier quoted context omitted.

Agreed. Where may I read about how to set up an LLM similar to that of Claude, which has the minimum length of Claude's context window, and what are the hardware requirements? I found Claude incredibly useful.

Looking into running models locally, maybe a 405B parameter model sounds like the place to start. Once understood you could practice with a private hosted llm (run your own model) to tweak and get it dialled in per hour, and then make the leap.

And now you can get the 405b quality in a 70b according to meta. Costs really come down massively with that. I wonder if it's really as good as they say though.

Re: ChatGPT Pro

#988

Post by wkat4242 »

paxys wrote: post #767
minimaxir wrote: post #41

The main difficulty when pricing a monthly subscription for "unlimited" usage of a product is the 1% of power users who use have extreme use of the product that can kill any profit margins for the product as a whole. Pricing ChatGPT Pro at $200/mo filters it to only power users/enterprise, and given the cost of the GPT-o1 API, it wouldn't surprise me if those power users burn through $200 worth of compute very, very…

$200 is a lot of compute. Amortized over say 3 years, that's a dedicated A100 GPU per user, or an H100 for every 3 users.

Not counting power or servers etc. But yeah it does put it into perspective.

Re: ChatGPT Pro

#989

Post by wkat4242 »

onlyrealcuzzo wrote: post #670
tokioyoyo wrote: post #16

Earlier quoted context omitted.

Qwen has an open reasoning model. If they keep up, and don’t get banned in the west “because security”, it’ll be fun to watch the LLM wars.

> and don’t get banned in the west “because security”, It's from Alibaba, which is Chinese, so it seems likely.

If it's an open model, good luck preventing us from downloading and using it though.

Re: ChatGPT Pro

#990

Post by xeckr »

replwoacause wrote: post #974
xeckr wrote: post #627

Never bought a $200 monthly subscription so fast in my life.

What’s your impression of it?

So far, I don't get the impression that o1 pro mode is even close to 10x better than GPT-4/4o, despite costing that much more. Definitely nowhere close to the kind of leap we saw from GPT-3 to GPT-4. It's good as a programming assistant, but waiting 1 minute+ for the output does interrupt my workflow somewhat. It also doesn't have access to the memory function or web browsing.
Post reply on HN