How many would be interested in Claude DHF subscriptions, for $20-$200 ones?
I liked the previous proposal by @deathwing and what it aimed to do, however, I'd like to see some more transparency this time around of everyone involved. So I'm proposing some changes to how things were to work.
- You have to share the repo with me so I can take a look at what you're working on to make sure it aligns with the ideology of the proposal.
- It not being open-source is okay as long as it's something that's building on Hive and is a working product (or can quickly enter MVP if you're just starting out).
- Payments will occur in HBD rather than fiat/gift subscriptions, I believe everyone knows how to withdraw hive at this point and I don't wanna complicate things with taxes/other personally (I'll also include VAT in the HBD amount). This also keeps things transparent as to where the funds are going.
From my perspective, I'll be judging the projects from a neutral perspective and using Claude to summarize them and take a look at activity, etc. I will also not share/leak any closed sourced projects/ideas with outsiders if you're using those to apply for DHF funding. Keep in mind however that they need to be directed towards Hive.
I'll also be adding a small fee for my activity and accounting as that may take time and effort which could be a reason not more people are stepping up to run this.
Either way, let me know what you think of doing things this way this time around and if you'd like to see that proposal be restarted as you deem it valuable to our users/devs. Or if you have any adjustments to the way it should be run you think would be better.
Update
Proposal is out.
Peakd: https://peakd.com/me/proposals/386
Ecency: https://ecency.com/proposals/386
Hive.blog: https://wallet.hive.blog/proposals
I think that if you cannot afford at least a $50 a month subscription to AI dev tools, then you probably are not serious about developing applications. So I don't think that funding Claude via DHF is a good idea. Especially since in my experience Claude is all marketing and is completely overrated and overpriced compared to GitHub Copilot and Codex.
Don't have xp with the others so can't really say, maybe this can work for the other AI's too, it's not really that important what you use more about what you're doing. and if I can create things I'm sure other non-devs can too, so it could empower some users to be creative and build some things people may wanna use.
Here is the basic comparison between the two, though it is not entirely accurate:
Level of autonomy is pretty much the same it also plans and tests and executes and commits in GitHub Copilot. And the context window in GitHub Copilot is at least 400K and actually depends on the LLM model Copilot is using so that is wrong too :)
Here are the cheaper models in GitHub Copilot that I like to use:
Here is the example of the 40 actions via GitHub Copilot or a day worth of building software and you can see that it was planning and doing other things, it also automatically suggests a cheaper LLM model which you can see it used GPT 5.3 codex and not GPT 5.5 or 5.4 which made working with it dirt cheap while producing perfect code:
I worked with Claude Code as well, but abandoned it because a session like the one above that lasted all day could easily eat up $40+ instead of eighty cents and the output is the same and sometimes way worse with more issues.
Well, I'm interested. I'm currently using Claude to develop two games, and I'm using a Pro version that I pay for out of my own pocket. I use this subscription for everything—whether it's for personal use, for work, or for my "hobby," which is game development, my current focus.
https://horizonforge.up.railway.app/
https://hivegrove.up.railway.app/
While I wouldn't personally submit anything at this point, I like the idea and love the accountability aspect of it.
I know there have been a lot of talks about slashing all funding, but building new things to attract people to Hive is worth the cost in my opinion.
!BBH
!PIZZA
!ALIVE
I am definitely interested and will share my main repos with you via dm :) I guess both are very well aligned with the target here. If I need to make a dedicated post for transparency I am down for that as well of course!
This sounds great Acid.... I think the first experiment was fine, but it lacked a way for us to see what fruits came from it. So if you can pull this off, you might actually create the blueprint.
:D
I wonder we can Claude some Tools for those got an Claude Subscription and can easy share their results ^.^ For me, i'm free to write a post at some point about what i have used it in the hive eco system but yea, dont saw that much results from those got a subscription
yeah, I guess I could give a summary of what I've had claude research while trying to be vague about close sourced projects to not give away their ideas, as long as it's proof that someone is working towards something that benefits hive and it can transparently be reported along with transactions going out of the proposal account. I think it'd be a much better way to do this.
I dig it, supported :)
I think it's valuable for the devs on hive, can help a lot... I didn't mind the previous proposal, but I don't know how many know what was built with it as of now, I know what you are doing as I follow, but others? It would be nice to have a monthly report with the developments done by each subscriber... So one can even use what's been done
I am a developer and I pay for all my own tools and have repeatedly over the years been told that nobody wants anything new without whale backing so how are vibe coded apps going to fare better?
Seems all we will end up with is another rally car. Money spent with nothing to show for it.
If there are metrics about growing Hive and milestones and repercurssions then I support it, if it is just a way for folks to play with hobby chatbots then no way.
I do plan on summarizing the things I can publicize on what people are working on.
What I can tell you is that I've built a lot more things than simple chatbots. Feels like you're undermining the skill of AI coding.
I'd say AI slop is undermining the value of ai coding because the last thing we want is more slop. Let's face it, 80% of folks using gen AI are because they have little motivation to work toward anything or put in effort, they want the end result minus the practice, learning, experience that it would take otherwise.
In my own posts I have said how I use these tools, and where I see they fit. I am not anti-ai, I am anti-slop.
We'd love to be a part of it! We'd like to continue supporting projects even if our latest marketing proposal isn't approved.
We have several projects in development: https://www.hivecreators.co/products
I would be interested, but maybe in a month or two.
Still enjoying my vacation from Hive, lol.
I am very interested, my pro sub is coming up in august, together with other free tiers I have found 20/mo to be enough.
Like you I have many different "things" I am working on. Is one repo enough to apply or what do you think?
!PIMP
!PIZZA
Sure, I will however look at activity as well so it needs to be a repo that's actively being worked on for Hive and not just used as a reason to work on everything else.
Let me share this, there are 6 repos linked here: https://github.com/srbde
I have been choosing to not vote for DHF proposals lately. This however feels it could be worth it for the community. I especially like the emphasis on transparency and reporting.
ill ask a few specific questions.
How many hive users do you plan to give the Claude pro account to? What will the selection/application process look like?
You say you will charge a fee for your time toward accounting and reporting. (Which sounds fair)
What will that fee be?
I'll try to evaluate each application neutrally and at the same time try and let applicants know if their idea/project already exists/is being worked on to attempt to save time/resources. I'm also going to have to keep some things private as to not spoil their ideas/ambitions from having other projects/users copy/steal them before they're built/announced.
I'm not entirely sure on the fee, I was thinking of checking the first couple months the workload/time/input of mine required, I'm guessing 10-15% but not higher than 20. I also plan on staking excessive HBD so we can make the most out of subscriptions.
For newcomers I expect them to show they understand what they're doing and upon renewals check that they're up to speed on security of the projects for end users.
Plus potentially more requirements/restrictions as we move forward and things are made apparent.
I support your proposal, wishing you much success.
Agree!
:)))
?
yeah thats what i do i make games insert coin to continue :)))
How are people devs and can't afford $20 for a Pro Claude Sub??
Thought Devs were paid well. :O
Worthwhile ideas can come from anyone, maybe some never bothered to check things out, this could be an opportunity for them to try.
I know I wish I had taken the plunge earlier than I did.
Fair enough.
I bought Claude Pro this month and it have been quite fun! :D
Start building something cool!
Got any ideas on what we need?
I asked on snap some days ago if anyone had an idea on what we needed, only one commented xD
Right now I am just automating shit, like power ups and such and a 2 games to see what Claude could do :D
$PIZZA slices delivered:
@ecoinstant(1/20) tipped @acidyo
Learn more at https://hive.pizza.
I am all in favour of any Generative AI being behind a paywall because it's not only opt-in, it also means that people will end up paying for their laziness.
All Generative AI should be behind a paywall IMHO. Might make people think twice about the "easy way".
I mean it kind of is, unless you download local model and make your gpu sweat.
Greetings and welcome to the new improvement-oriented projects. If this involves benefits, control, and oversight, and the goal is to improve Hive, you have my full support.
The proposal sounds great; it inspires a lot of confidence in achieving the goal. If it's aimed at generating peace of mind and better results, I'll be there to try it.