Ask HN: Should I subscribe to ChatGPT Plus if we can get it for free on Bing? | Hacker News Reader

Ask HN: Should I subscribe to ChatGPT Plus if we can get it for free on Bing?

75 pointshackerlight88 comments

Apparently I can get access to GPT-4 for free through Bing[1]. So is it worth it to pay $20 for GPT-4 for personal use? Is the API really that much better than using it through Bing?

[1] https://www.kdnuggets.com/2023/05/3-ways-access-gpt4-free.html

View on HN

> is it worth it to pay $20 [...]?

I always object when someone tells me "it's (only) X $ / Y $" and I know for me the relevant part that I want to say out loud is per month, as in, it's a subscription, not a one-time or prepaid amount. I've rid myself of all subscriptions to the degree feasible because whether it's 'little' or 'much' it is recurring, per month so there will likely come a point in the future where I'll just pay without really using that thing anymore or where I forget or can't pay or where I have to pay although that sum could better be spent elsewhere. This is my experience with all subscriptions. I'm even getting my cooking gas as 'prepaid' (i.e. I buy bottles) and I wish I could do the same with electricity and so on.

So I'm totally fine with getting a gift card or whatever you call those thingies in the supermarket where you can throw a lump (20€, 50€) at some provider (Google Play, whatever) and then get to eat up that amount (over the course of at least one year, hopefully). By comparison, I wouldn't sign up even for 1€/month even tho I might get the same for 'less' money—it won't be less, and when you listen to Marie Myers (HP CFO, [1]), corporate knows darn well this is so.

* [1](https://www.theregister.com/2023/12/04/hp_printer_lockin/)

“Never subscribing” is of course one approach. Another would be “know your exits and keep a budget”.

Yeah, I feel "I never subscribe because I forget about the fact that I subscribed" is a statement indicative of a different problem.

Really though? When even toothbrushes want recurring revenue, it’s so easy to rack up services. I feel like I’m more conventional than OP but still avoid monthly subscriptions if at all possible (though pay for ChatGPT and other things that are worth it for me)

I'm not saying pay for everx nickle&diming sub, which I agree have gotten egregious. But for the ones that you do subscribe to, write them down so you don't forget about them.

To add to this excellent point, a better way to compare and understand costs in our subscription-addicted world is to annualize them. So with ChatGPT we know we'll be on the hook for $240/year for this new thing and can compare that number to all manner of things to decide whether it's worth it. (This also gets you thinking about annual discounts and what you can negotiate in the long term, so it can unlock free money.)

An EVEN better way, albeit more complex, is to amortize costs for comparison. This basically means figuring out what the entire cost of the thing is from beginning to end and then using that as a comparison to other stuff. This is where subscriptions get us - 15 years in, the amount of money I've paid to Netflix (even as the quality of the programming seems to have eroded) feels a little crazy.

Thinking about amortized costs when you have a little bit of cash in the bank or access to cheap credit is when finance starts to get interesting. Let's say you know you're going to need something for the next 10 years. Instead of paying the regular monthly price, what kind of discount can you get if you offer to pay cash up front for 1 year or 10 years? You'd be surprised how much people can be willing to lower the price of something for a large enough chunk of cash up front. For example if you live in a weak property market this can be a good way to rent for very cheap.

At the very least, the end of the year is a great time to look back at what you spent and ask, am I happy with spending $XXX or $XXXX on that again in 2024?

I just model subscriptions as a perpetuity at 3% interest (i.e. treat it as a permanent expense pretending I'm doing retirement planning). Then for a monthly expense $x, it costs 400 * $x in capital to "buy" it. "Multiply all monthly subscriptions by 400" is an easy enough rule of thumb to remember.

Is Netflix worth $20/mo becomes equivalent to is Netflix worth $8000, and you can just examine the alternative ways you could use that capital (e.g. buying a few thousand used movies and building your own collection which you can share with friends and family instead).

Naturally I have very few "subscriptions". I do back up data to glacier for now, but plan to switch that to off-site back ups at my mom's/MIL's.

Do you account for the fact that you can "sell" the depreciated asset by canceling the subscription?

+1 on the rent point. I've gotten significant discounts on rent in the past by paying a full year in advance. This makes sense from the landlord's perspective too, as they're not only gaining the time value of money, but also eliminating the possibility of missed payments.

Well, it's known that it's a subscription. I am not sure what you what, exactly.

I want the explicit mention of it. I realize it's rather younger than older folks who give subscription prices as a sum instead of sum per time, and incidentally internet speed in terms of (Giga-) bit instead of bit per time. Maybe it's just me or the pedants who were responsible for my education, but "it's known it's a subscription" is fairly weak; what tells me this is a monthly, not a bi-monthly, weekly, yearly or one-off price? Listening in on the conversation of others in this and other threads I also gained the impression that there are some providers who do charge for some of their neighboring services an extra volume-oriented fee, so it's far from being that simple in the end.

I don't think the subscription fee in itself is the problem, but the lock-in and inevitable enshittification. I would not trust "OpenAI" to be any mandatory part of my workflow. It is really hard for me to change my workflow, since I need it in more or less muscle memory to have any stamina in doing stuff.

> I wish I could do the same with electricity and so on

If you are using e.g. wood pellets for heating you essentially can.

> wood pellets

or solar but I can't do that where I live

> I'm even getting my cooking gas as 'prepaid' (i.e. I buy bottles) and I wish I could do the same with electricity and so on.

Gas and electricity do have small fixed components (“hookup fees”) but for the most part these aren’t strictly subscriptions since you pay for what you use.

> So I'm totally fine with getting a gift card or whatever you call those thingies in the supermarket where you can throw a lump (20€, 50€) at some provider (Google Play, whatever) and then get to eat up that amount (over the course of at least one year, hopefully).

So you are okay with subscriptions as long as they won’t charge you more than X? You might like privacy.com. They let you create credit cards with unique codes and defined spending limits. You don’t need to worry about cancelling anything since your billing just won’t go through after a certain amount is spent.

> So you are okay with subscriptions as long as they won’t charge you more than X?

No, most definitely not. I'm not talking about the sum (although we have to talk about it too), I'm talking about the recurrent obligation.

> for the most part these aren’t strictly subscriptions since you pay for what you use

which is where the obligation comes in. I prefer setups where I show up with a sum of cash in my hands and I can go home and use up whatever I bought. The utilities have an idea about what I might be using going forward so they set a price point for me to pay up for; each year there's a round of reckoning where I get back money or have to pay extra (Ukraine War -> both gas and electricity get more expensive), and they set a new price point for me. I cannot influence that in any meaningful way; should prices soar, I'm left with swallowing it or terminate the contract altogether. I could live by candle light (in theory) and use computers with a hand crank and battery setup (in theory), but I cannot save energy now and be the master of my expenses, now. When I was a kid we did have an electricity gauge with a coin slot in the vacation apartment but that's fifty years ago. Something like that would be great. I manage to use my phone like that, and my cooking gas; now that we all must have smart meters, I want the same for electricity too. Toss a coin, get some light, stay out of subscription, great.

You can 100% do this in many states. But solar, create your own energy and the state will force the electricity company to buy the excess capacity away from you.

I don't live in the states and it's not a regulatory but a technical problem, living in a rented flat that has hardly a good spot for a PV panel to hang to...

I haven't used it through Bing recently, but last I tried, I found whatever custom prompt and/or tuning they're using made it much more cumbersome than chatgpt plus. It's far less willing to give you answers from training data, and instead wants to search the web for everything. Sometimes that's useful, when you want it to search a website for a particular piece of information, say. But most of the time I can more effectively search the web myself. Where I find gpt-4 most useful is when it can give me answers, or at least a starting point, directly from its "knowledge". (Which I would of course verify as necessary.)

Like I said though, I haven't used it recently. I've added a custom prompt to chatgpt to only search the web when necessary, if it can't provide a good answer otherwise; if you can now do something similar with Bing chat, and it respects it, maybe it will work fine. (Though when I did try it a while back, I found it more restricted in general; the web searching was just the most obvious thing.)

I didn't use Bing chat that much, but everytime I've added "do not search" to the prompt it respected it.

In my experience, what you get on Bing is quite different from what you get on ChatGTP.

I would prefer the free ChatGPT 3.5 over what I get on Bing because Bing counts down the responses and tends to give much shorter answers.

Here is an example for Bing Chat vs ChatGPT Plus: https://dropover.cloud/4e1dd7

If you don't plan to use it enough to worth $20 a month, you can purchase some API credits from OpenAI which don't expire and use it directly from their developer playground.

Actually, IIRC they will give you some free credit valid for some time when you create an account on their developer platform so you should be able to give it a try yourself.

Go to: https://platform.openai.com

Create a new free account, then head to the http://platform.openai.com/playground and try out ChatGPT with GPT-4(might be unavailable on free credit until you top-up some money) on their Chat interface.

I seem to only be given a selection of 3.5 type models if following those steps.

Your screen should look like this and you should be able to select GPT-4 from the menu: https://dropovercl.s3.amazonaws.com/c4606813-cbc5-4f66-b4f7-...

If the GPT-4 options don't appear, maybe they don't give them to non-paying users? If that's the case, you can add something like 5$ credit to your account.

You might have to top up with a few $ of prepaid credits before you get access for the first time.

https://help.openai.com/en/articles/8264644-what-is-prepaid-...

This is it. I don't know why they hide models if you don't have enough credits.

Probably supposed to prevent abuse (people cresting bot accounts to use the free credits with GPT-4)

> you can purchase some API credits from OpenAI which don't expire and use it directly from their developer playground.

Or via https://github.com/yakGPT/yakGPT

I didn't use Bing much but in my experience, it's a lot harder to achieve what I want. For example, it almost always search the web for an answer which is a hit or miss for me.

If you want to use GPTs or to generate images with Dall-E then ChatGPT Plus is a no brainer I guess.

But if your focus is on GPT-4 then I highly recommend to use an API instead.

There are multiple pros of using an API Key:

- You pay for your usages. I've been using API Key exclusively and most of the time, it cost me around $5-$10 a month

- Your data is not used for training. This is important for a privacy-minded user like me. Though you can disable this in ChatGPT (but you will lose chat history)

- No message limit. Though there are rate limits to prevent abuse but generally, you do not have the message limit like the ChatGPT

- You can choose previous GPT models. In my experience, `gpt-4-0314` was the peak gpt-4 model.

Depends on the applications, you can also get these:

- Access to multiple AI services: OpenAI, Azure or OpenRouter

- Build custom AI workflows

- Voice search & text-to-speech etc.

- Deeper integrations with other apps & services

There are also a few cons:

- No GPTs support yet

- If you use Dall E a lot then ChatGPT plus is more affordable. Generating images using Dall E API can be quite expensive

Edit: Some tips when using an API Key:

- You pay for tokens used (basically how long your question & AI's answers). The price per chat message is not expensive, but usually you will need to send the whole conversation to OpenAI, which makes it expensive. Make sure to pick an app that allows you to limit the chat context.

- Don't use GPT-4 for tasks that doesn't require deeper reasoning capabilities. I find that GPT-3.5 turbo is still very good at simple tasks like: grammar fixes, improve your writing, translations ...

- Use different system prompts for different tasks: for example, I have a special prompt for coding tasks, and a different system prompt for writing tasks. It usually give a much better result.

Shameless plug: I've been building a native ChatGPT app for Mac called BoltAI[0], give it a try

[0]: https://boltai.com

kudos for perpetual license. app looks neat. non-techy question: how do you ensure users their keys are safe?

Thanks. By default, BoltAI stores user's API key in Apple Keychain. So other applications and scripts won't be able to access this key without triggering the OS's security dialog.

Secondly, all requests are sent directly to OpenAI servers. There is absolutely no middle/proxy server.

There is however a proxy server for Claude but this was due to a technical difficulty and user can host their own proxy server instead.

thank you for the clear explanation

It is not a perpetual license. "One year of free updates included." means subscription. Especially considering product depends on third party API, which rapidly evolves, you cannot really stick with the old version, because some new model will not be embedded in the one year old version of the app you have. So, in reality you pay $40-80 a year for this product. Plus your ChatGPT API Key usage. In that case I would just suggest pay for the OpenAI.

If you are looking for an alternative of the app for the ChatGPT - I would suggest https://www.typingmind.com I have been using it since it was released. It is also available with Setapp subscription, which gives you some free usage of the OpenAI with Setapp embedded API key.

Technically it's perpetual fallback license, which was popularized by the Jetbrain team [0]

> A perpetual fallback license is a license that allows you to use a specific version of software without an active subscription for it. The license also includes all bugfix updates, more specifically in X.Y.Z version all Z releases are included.

After 1 year, you can continue to use the app without having to upgrade. You won't receive new features, but will continue to receive bug fixes for the features released previously.

So I can argue it's not "subscription".

Plus, renewal cost 50% less, so it's not $40-$80 a year, but only $20-$40 a year after the first year.

And lastly, the app fetch model lists from OpenAI API so there is no need to "support new models".

I believe this is a sustainable pricing model for an indie developer like me. And I believe it's fair for everyone.

[0]: https://sales.jetbrains.com/hc/en-gb/articles/207240845-What...

I feel like if you have to specify “technically”, you agree with me that it is basically a subscription.

As an Indie developer myself I do not support subscription based licensing (or your version of license) for B2C products. You can offer B2B subscription based licensing, but B2C should always be perpetual licenses. Sure it is ok to ask for the upgrade to new version, which has significant features.

And to clarify, we are talking about the product that does not have any outgoing server spending.

That is my opinion.

> I feel like if you have to specify “technically”, you agree with me that it is basically a subscription.

No, I am not. The "technical" part here is the "fallback" in "perpetual fallback license".

Appreciate your opinion and I'm sure many think the same way.

But selling perpetual licenses just won't work for me. It's not sustainable, especially for this kind of app.

that's no different than pay the full price for new version that comes out every year. do you remember how perpetual licenses use to work?

I feel like Sublime Text showed the best of it. You purchase the new license every new major version. Affinity apps are the same. Acorn. Alfred. And a lot more. There are certainly some apps that release a major update every year. But in that case you pay for work already done. In case if you pay for subscription, you are going to see minor updates “bug fixes”.

I don’t see a reason for an indie developer to charge subscription, if there is no server cost. New versions should just make the app interesting for more users, so the income should be flowing constantly every month.

Anyway, I just personally don’t support subscription based apps and never will.

Is there a relationship between MindMac, which I have seen before on HN, and BoltAI? Websites are quite similar.

No, there isn't. If you're talking about the design then I can see both were modified from a paid Tailwind UI template. For marketing copy, all I can say is they were "inspired" by mine. You can easily check using Wayback Machine [0][1] :D

[0]: https://web.archive.org/web/20230522013022/https://boltai.ap...

[1]: https://web.archive.org/web/20230711214004/https://mindmac.a...

I think your section on data privacy could mislead some folks. You’ve written:

> Is my data protected? Absolutely. BoltAI operates locally on your device, and no user input or prompts are stored or sent to us.

While the explanation might be true, as far as I understand it, you can’t assert that users’ data is “absolutely” protected, because it fundamentally depends on OpenAI to keep your users’ data safe too, which you can’t control.

I would guess most of your audience would understand this anyway, but I would be careful as potentially someone could sue you in case OpenAI had a data breach or something.

I think you're right. Thanks for the feedback. I will update it soon!

This is the way - I just had chatgpt make me a personal use command line utility that just uses the API to mimic chatgpt with whatever model you pass in, saves the running chat as a text file, and gives you a running cost with every prompt. It took about an hour to get something I was happy with, but now I average about a buck a month on it (but I still use the free chatgpt/claude for simpler queries)

> But if your focus is on GPT-4 then I highly recommend to use an API instead.

When you have made that choice is it still possible to use 'browse the web' or 'talk to a PDF' like functionality? If so, what would be the best way to do it?

It depends a lot on the GPT client (third-party application).

My app (BoltAI) does support web search (via Google Search) and web browsing (built-in) but not chat w/ PDF functionality.

I built another app[0] specifically for that use case.

So I guess if you need an all-in-one kind of functionality then ChatGPT could be a better choice.

[0]: https://pdfpals.com

I'm a big fan of TypingMind as a front end. typingmind.com

Haha yeah, same. TypingMind is probably the client with the most features. Tony ships at a crazy speed. With BoltAI, I want to focus on MacOS with more native integrations.

Heh, yeah, he is a crazy coder no doubt. Good to have a Mac specific option though. Good luck with it.

>> In my experience, `gpt-4-0314` was the peak gpt-4 model.

That's interesting. Why do you say that? And where can we access the model from?

It’s just my personal experience. So I mostly use GPT-4 for coding tasks and gpt-4-0314 usually gives me the answer I want, without having to use system prompt tricks like this https://twitter.com/voooooogel/status/1730726744314069190

If you use API Key then you can change the model parameter. Here is the list of models that OpenAI support: https://platform.openai.com/docs/models/gpt-4-and-gpt-4-turb...

Too bad it’s considered legacy model and will be deprecated next year.

Important note: Bing balanced mode (default) uses GPT 3.5

Only Precise and Creative modes use GPT-4

https://twitter.com/emollick/status/1732495030143549541

Also see:

An Opinionated Guide to Which AI to Use: ChatGPT Anniversary Edition

https://www.oneusefulthing.org/p/an-opinionated-guide-to-whi...

I actually find Bing superficial and impatient compared to ChatGPT-4-Turbo, and way more inclined to give partial or incomplete answers, and jump to unsubstantiated conclusions and give more “hot takes” where chatGPT is way more circumspect. Bing is oddly more likely to hallucinate, and even give hallucinated links, which fortunately the UI strips out.

Bing really feels like a different GPT 4 — tuned for brevity and to reduce resource utilization. (Except for images. Bing wants to generate images for no reason, while chatGPT sometimes double checks to make sure you really want one, and has scaled back the number of simultaneous variants it generated from 4 to 1, although no limit that I’ve reached yet.)

Bing has a max turns per chat limit I think around 20 or 25 at the moment, at least with the purple (formerly “Creative”) GPT 4 model. ChatGPT doesn’t have conversation limits, although older text falls out of context. Bing created that limit to prevent the model from going unhinged after too many turns, which happened frequently with Bing but almost never with ChatGPT for some reason.

I’ve also hit a daily limit on Bing while not on chatGPT+ — chatGPT4 currently has a 30 turn limit within 3 hours, although I’ve rarely hit that outside of long voice chats. I do try to cram multiple questions into one go, to conserve turns though. Or I give certain tasks like summaries to chatgpt3.5 as it’s faster with no quota.

Plugins were briefly interesting. Maybe the new custom GPTs will become a must, but not yet. Oh, and the python runtime environment is epic. It can literally build its own tools to answer questions.

If I couldn’t afford $20 a month, I could maybe make do with Bing, but if I could afford it, I think chatGPT is very worth it.

You mentioned the API, which is a whole different thing than the $20/mo chatGPT subscription. It’s usage based, and designed to allow you to integrate GPT into other software and services. It has different models and different pricing tiers. I use both chatGPT and the API, though the API I use only lightly for personal use and it hasn’t consumed my first $5 deposit yet. There are also chatGPT like third party clients that take an API key and act like chatGPT and depending on your usage may cost less or more than the normal chatGPT subscription.

It is believed that the characteristics you mention in the first few paragraphs are because Bing is based on an old pre-release version of GPT-4, which didn't have all the guardrails that the final versions have (note that Bing Chat was released before access to GPT-4 was available directly from OpenAI).

I currently pay because I use it a lot on my iPad and phone in addition to the web. Anyone have a recommendation on how to use api access on mobile? Then changing might be worth it

Use an alternative frontend like Chatbot-ui or takeoffchat. They let you use your own api key

Anyone know of a good Assistant API frontend yet?

The author of LibreChat [0] teaserd on discord that he's working on one. LibreChat is already the best "chatgpt ui" clone I could find (it has some rough edges, but not as many as the official chatgpt ui).

[0] https://github.com/danny-avila/librechat

Was also recently discussed here https://news.ycombinator.com/item?id=38502805

Spin up a temporary credit card using a service like Wise, pay for 1 month, and then close the card. If you're like me you'll switch it to your normal debit card within a week.

$20 a month is enough of a commitment in the developed world to sting just a bit on something you're not even sure you'll use much - but $20 once, as an exploratory measure, is a much easier proposition to justify.

Ever since (custom) GPTs got released last month, GPT-4 has become pretty much unusable around 2/3 of the time, performance-wise. Tons of network errors.

So, no, it's no longer worth it. At least not until they free up resources for paying customers by moving those useless, resource-consuming toys somewhere else.

For me it has been fastest ever for the past few days. You don't need to wait forever for text generation to finish anymore.

I’d say just try it for a month and see. Only you can tell if it is worth for you. In this case being able to subscribe for a month only (compare to others which might have a much more expensive per month fee if not paid annually) is an advantage you should take.

What you said is exactly what I thought. And above is what I did and I’m glad I made that experiment (of trying ChatGPT Plus.)

By the end of that subscription month, the question I asked is if it helps me enough to save me time worth more than $20/months. In that sense unlike other comments, I don’t care if it is recurring cost, in any month that its effectiveness didn’t shave me $20 worth of time then I’ll unsubscribe it.

I pay $21 a month (after tax) for GPT Plus because Bing limits me to one prompt without followups; I can not carry on a conversation. GPT Plus lets me respond to what the AI told me.

Nigh-exact duplicate thread from 22 days ago: https://news.ycombinator.com/item?id=38316790

The custom GPTs are convenient. Instead of digging through my collection of prompts for different tasks, I just click one of my custom GPTs that has been preinstructed.

Neither, use Bard, because both OpenAI and Microsoft have Customer Noncompete Clauses in their Terms of Use.

Is Github Copilot Chat using GPT4 or GPT3.5? I already have Copilot Chat trough Copilot subscription.

It’s using 4 turbo but it’s through a custom assistant

If service is hidden behind a login, it's probably not free at all. Of course, nowadays no one can guarantee that even after paying subscription your personal data won't be used in some peculiar ways.

I pay for ChatGPT to get the option to avoid the promt history and training on my promts. If there is a way to do this for free, I'd love to know about it. The article does not seem to take it into account.

On iOS open ChatGPT app -> settings -> data controls -> disable chat history & training toggle

It’s something similar on the web

https://privacy.openai.com/policies "do not train on my data request"

Bing's responses are not at all same as GPT-4. I don't know the exact reason but they are definitely not using the same model or base instructions. GPT-4's responses are far better.

What you should do is use a local application like MindMac or similar and use GPT-4 via the API.

It's a _lot_ cheaper unless you use it really heavily. I've yet to break $10/month (+taxes)

Very polished. Is there anything similar for windows?

There are many free GPT frontends, especially web frontends, e.g. https://github.com/yakGPT/yakGPT

“Anything similar for Windows” would mean a native application, like MindMac is on macOS.

Does this support the new tailored "GPTs" that just got released?

There's also an implementation that works as a Discord bot: https://github.com/Zero6992/chatGPT-discord-bot

You host the bot yourself so nobody else can see what you're doing.

The pricing looks rather steep. Does it bring any benefits relative to just using openai playground?

It can access multiple different AI APIs, Open AI and Azure AI just among them. You can even use local models with it.

I can also bring it to the front with a hotkey and start typing a question, it's easy to switch between GPT3.5, 4 and Dall-E. You can also give it different pre-prompts from a quick menu etc.

I found the Playground and the OpenAI web UI just janky enough so that I started looking for a proper application.

Got it on Black Friday sale, when it was 55% off or something.

Thanks. Sounds good, really - but I'll have to wait for the next Black Friday deal.

It’s not 1:1 the same thing though. Gpt4 chat does web search for research, tries to run python code etc

The API is pay per use. $1 minimum spend before you can use gpt4 and they take a while to bill you.

It’s a separate thing to the $20/month chatgpt plus subscription.

Bing tends to stop the conversation when it receives harsh criticism. ChatGPT will be sorry and try to do better.

Use the API and find a client that works for you. I find I spend way less than 20 per month.

Do you need api access? If so you need to either pay Microsoft (azure) or OpenAI

now you see why openAI's had to go through the whole board Vs sam altman showdown

Yes.

Get the API and a client like MacGPT.

No. Use your own mind.

Reply on news.ycombinator.com