I pay $20 for ChatGPT, and I can use it on OpenCode without paying extra. I haven't tested Claude. Is it much more powerful?
I'm not much of a coder; I'm more on the PM side of things, but maybe it would be interesting for the devs in @mantequilla-soft and @aliento.
I think that the proposal is great, but it's bad timing, I guess.
Why not create the fund some other way? Why use the DHF?
Claude Code completely destroys ChatGPT for programming. I had this same kind of discussion with a bunch of my programmers at the beginning. They were using all kinds of AI products (other than Claude Code) and I basically had to force them all to switch. But they've all seen the results now.
Another point a lot of people don't get: Claude (the LLM) and Claude Code are different things. A bunch of my PeerVerity team programmers were using Claude with Github Copilot licenses and only paying $10/month, so they were also wondering why I was so forcefully pushing a $200/month license. But Claude with Claude Code performs at an entirely different level, because of the instrumentation around it provides.
Thank you for the explanation!
Makes more sense now.
We briefly discussed it in our latest report on @mantequilla-soft, fast-forward to minute 46:30.
We agree that it's a great opportunity to develop faster, and development budgets should be smaller. I think we will apply, maybe not me directly, but @meno and @kesolink are considering it.
I think you have not used Codex CLI recently, it has almost all the instrumentation that Claude Code does. (skills, MCP etc.)
You should give using both a go, hell give your Claude Code, Codex CLI agents. You'll see what I mean.
These things advance so fast that it's hard to keep track of everything.
Similar, depending on the situation at the time (compute, model quality etc.) one may be better than the other. Harnesses (like OpenCode) are regularly updated too which may result in different results even if using the same exact model.
Out of curiosity, how?
Author rewards, beneficiary rewards, witness rewards, there are many ways. Using the reward pool in general, that way we don't bring in extra inflation.
Many whales are mainly voting on things like the @hbdstabilizer, and it's getting a lot of rewards.
This is for the last few weeks:
https://peakd.com/@hbdstabilizer/activities
But then you're forcing one person (or a few persons) to pay for the project, technically, this is exactly what DHF is for.
I don't think hbdstabilizer is a good example though, all rewards (bonuses on top) hbdstabilizer receives goes back to the DHF anyways.
But this way, people can actually downvote if they do not agree.
Yeah, maybe not the best example, but my point is that there is enough power, even at this valuation, to cover that and more.
I think I will make a post to give more context.
To the same argument, people can simply not vote for the proposal if they disagree.
I'll get more feedback from the community before deciding if I will support or not.
And will talk with the Mantequilla Soft guys about it too.