Live data from Hacker News

Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users

openai.com

261–270 of 280 posts

Re: Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users

#261

Earlier quoted context omitted.

Claude's rate limits are absolutely shit for free users. 3 chats and its gone for 24 hours.

It's one thing to be a free user...quite another to be an entitled free user. This is precisely why I do not offer anything for free. I've been there and done it back in the iPhone 3 era with a dozen apps. Learned my lesson quickly. Never again. It was a nightmare. If the product isn't good enough for someone to pay for it (that could be direct user-paid or advertiser supported) it should not exist.

If your free tier is marketed as a free tier but really it's just "your first hit is free" like your crack dealer's marketing, then it's a fraud on the user, a scam to get you hooked and paying. I am not an entitled user for thinking a free tier designed to provide half the answer and make me pay for the second half is a scam from an evil corporation and to call that out.

Also, it is intended to be supported by ads, so it fits your ideal product description quite well despite being a scam.

Re: Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users

#262
post #7

I think it’s a misread to think the default ChatGPT model switching to GPT 5.6 Luna is some sort of desperation move. Keep in mind that Claude .ai never had this extreme stratification between the frontier models and the free tier (Sonnet is available to free users with rate limits). So 5.6 Luna is just their next version of what they used to call 5.5 instant tier And 5.x instant models were never much to write home…

Honestly, given the price 5.6 Luna is pretty fantastic. I replaced GPT 4.1 mini in several projects. It's also 50% cheaper if you only need it to run sometime in the next 24 hours.

yep, agreed. little bit smarter than deepseek 4 flash, but a little bit worse at tool calls.

Re: Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users

#263

Earlier quoted context omitted.

> but it would just sound dumb or biased We have one of those: Grok.

Yes, and it's indistinguishable from the average X user.

Bold of you to assume they’re actual users.

Re: Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users

#264

Earlier quoted context omitted.

It's one thing to be a free user...quite another to be an entitled free user. This is precisely why I do not offer anything for free. I've been there and done it back in the iPhone 3 era with a dozen apps. Learned my lesson quickly. Never again. It was a nightmare. If the product isn't good enough for someone to pay for it (that could be direct user-paid or advertiser supported) it should not exist.

If your free tier is marketed as a free tier but really it's just "your first hit is free" like your crack dealer's marketing, then it's a fraud on the user, a scam to get you hooked and paying. I am not an entitled user for thinking a free tier designed to provide half the answer and make me pay for the second half is a scam from an evil corporation and to call that out. Also, it is intended to be supported by ads,…

> "your first hit is free" like your crack dealer's marketing

Oh, please.

Don't use it then. Simple.

Sign of entitlement: You get something of value for free, bitch about limitations and make-up bogus reasons to continue using the product without paying.

If it is of value, use it and pay for it. If it isn't, do not use it and you have nothing to worry about. You are bitching about having access to untold billions of dollars of infrastructure and a massive amount of work from people who have lives and families FOR FREE. Get a grip dude.

I use it for work, extensively, and I gladly pay for the services from a number of AI providers. I would gladly pay double. It's the no-brainer of the century: I can do months of work in weeks or less for a few hundred to a few thousand dollars per month.

I was talking to someone the other day who shares their Netflix account for four other families. Same thing, although, in this case, it's not entitlement, it's theft.

They are both versions of the tragedy of the commons in various ways.

Re: Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users

#266

Earlier quoted context omitted.

If your free tier is marketed as a free tier but really it's just "your first hit is free" like your crack dealer's marketing, then it's a fraud on the user, a scam to get you hooked and paying. I am not an entitled user for thinking a free tier designed to provide half the answer and make me pay for the second half is a scam from an evil corporation and to call that out. Also, it is intended to be supported by ads,…

> "your first hit is free" like your crack dealer's marketing Oh, please. Don't use it then. Simple. Sign of entitlement: You get something of value for free, bitch about limitations and make-up bogus reasons to continue using the product without paying. If it is of value, use it and pay for it. If it isn't, do not use it and you have nothing to worry about. You are bitching about having access to untold billions of…

Most of the stuff I use is free and excellent. Linux, Firefox, libreoffice, xcp-ng, bottles, steam, playnite, opnsense, docker, dockhand, immich, freshrss, obsidian, tailscale, calibre, caddy, docker, pi-hole, signal, audiobookshelf I could probably keep going for a long while. I'm very glad most good software creators I've come across do not have your mindset.

Re: Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users

#267

Earlier quoted context omitted.

How are the models too dumb? How are they dumber than the average person? I wonder if anyone gave Claude an IQ test (the one for humans).

I think it's Karpathy who coined the term “jagged intelligence”. LLMs are both extraordinary smart in domain they have been explicitly trained on (like Math) and positively dumb on things they haven't.

I'm that way too. So are most people.

That's why you shouldn't listen to your pop celebrities for political advice.

Re: Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users

#268

Earlier quoted context omitted.

> "your first hit is free" like your crack dealer's marketing Oh, please. Don't use it then. Simple. Sign of entitlement: You get something of value for free, bitch about limitations and make-up bogus reasons to continue using the product without paying. If it is of value, use it and pay for it. If it isn't, do not use it and you have nothing to worry about. You are bitching about having access to untold billions of…

Most of the stuff I use is free and excellent. Linux, Firefox, libreoffice, xcp-ng, bottles, steam, playnite, opnsense, docker, dockhand, immich, freshrss, obsidian, tailscale, calibre, caddy, docker, pi-hole, signal, audiobookshelf I could probably keep going for a long while. I'm very glad most good software creators I've come across do not have your mindset.

None of these things are free my friend. They probably represent tens of billions of dollars in development over years, decades in some cases.

You are missing the point entirely. Let's just take one: LibreOffice. This was a FOSS project from the very start. It is intended to be used as you use it. Even then, bitching about functionality and features would be entitled. In that case, if you don't like it, stop using it or participate in the development...for free, donating your time.

ChatGPT is NOT a free product. Yes, they have a free tier. And that's fantastic. We've all used it. But this is not a FOSS company. Bitching about their free tier isn't right. Again, that's being entitled. Why do they owe you more than they are giving you? They do not. If you want more, pay for it.

Or, you could buy a bunch of powerful computers and run an open weights model on your own hardware and have free intelligence to use as much as you want.

Oh, wait, that wouldn't be free, would it? You'd have to pay thousands for the hardware.

OK, wait, how about you do that and I use your self-hosted LLM's for free and then bitch about the lack of performance.

Yup. Entitlement. Always a reason for other people to owe the entitled and never a reason for the entitled to owe them anything. And this goes way beyond what we are talking about here, far and wide.

> I'm very glad most good software creators I've come across do not have your mindset.

Here's something else you do not understand. FOSS isn't free to create, not even close. Take something like Linux. I would be comfortable estimating that Linux has probably a development cost in the one-hundred billion dollar range. I would not be surprised if was way more than that. The massive number of people who worked on Linux over the decades have lives, families, homes, responsibilities, etc. The only way Linux happened is because they are otherwise employed and donate their time (with is fantastic, I did a lot of of that when I was younger, had time and less responsibilities) or someone is explicitly paying for them to work on Linux because of the benefits they might derive from it. The only "free" part about FOSS is that users of the software don't pay for it. Don't confuse that with all software having to function this way. Like I said, OpenAI is a business. It isn't your personal charity. If you want more, pay for it.

Re: Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users

#269
post #191

Earlier quoted context omitted.

That’s his definition and in no way a universal one. In the meantime, it’s still very easy to differentiate between an AI and a human in a chat. You just need to know the quirks of these systems. Like counting letters, hitting the safeguards, etc. So call me when one of them can pass the Turing test against me and then we can talk about AGI

> So call me when one of them can pass the Turing test against me and then we can talk about AGI Ring ring I'm calling you right now. We blew past the Turing test goalpost over a year ago, using 2024 models. https://www.ie.edu/uncover-ie/has-ai-passed-the-turing-test-... GPT-4.5 passed the Turing Test with a 73% human rating , outscoring actual human subjects. That is, human evaluators considered the AI more human th…

I'm not repeating myself

Re: Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users

#270
post #71

Earlier quoted context omitted.

Ads work when you can server $.01 worth of ads to a user for $.0001 of server cost. I fail to see how you can make it work when doing LLM inference which is significantly costlier than web search.

> I fail to see how you can make it work when doing LLM inference which is significantly costlier than web search. The median LLM query isn't significantly costlier than web search.

I don't know where you get that from, but it's pretty easy to prove it wrong: with their daily active users count, OpenAI and Anthropic would have been profitable for years if that was true. Instead they are bleeding cash like no others and are still trying to secure another round of funding 4 years after becoming among the most heavily used apps in existence.

Whatever you think about Altman and Amodei, they aren't clueless idiots and if ads was all they needed they would have done that from the beginning.

Post reply on HN