Please feel free to let me know any additional details or amendments needed.
I have done a very successful Twitch stream, 4 hours, in which I developed some R scripts for making a naive Signals submission in R. I have published the code. I posted the video to youtube. With some helpful feedback from a fellow numerati (thank you @filipstefano !) I have fixed the audio panning and volume spiking issues in OBS, so these should not be a concern. I promote my streams via my twitter account.
I would like to propose that the council incentivize me to do more of these streams. I would like to make my initial request for 4 streams, of ~4 hours each, approximately weekly, at a rate of 2 NMR / hour. So total, 32 NMR.
I would continue posting code developed on stream to my github, and recordings of streams to youtube.
At the time of this post, 1 NMR is approximately $30.32 .
2 NMR $60.64
32 NMR $970.24
I’m just hitting my stride in the tourney, so my interest in Signals is only tentative for now.
What could get me more interested in Signals, is if there was a way to use it as much like the tourney as possible. For me, that would mean:
* no "domain expertise" of stock market dynamics needed (or even used),
* a free (or very cheap) data set I can use in a "feature symmetric" way (as I do in the tourney).
Do you think this is possible, or is Signals just fundamentally different?
i think it’s less likely to achieve good return with the former requirement, but idk!
something i’ve already developed (and released, i think!) is a scraping program for yahoo stock data which is free. expanding on this may indeed be on my plate. also, having a script others can adapt may help bridge the gap re: your first point.
Do you have suggestions for targets for engagement that we could use as rough thresholds to determine if we keep going and going with it? How many people would you want to watch the videos or to download the code to make you feel like it’s worthwhile?
great questions, and I’m open to feedback on these (and other) metrics. This weekend i’ll review the figures on my existing material and post some goals here.
for the sake of completeness i wanted to state that although for this proposal, i’ve proposed hours broadcasted, this work will of course entail me doing work before and after a live cast, this amount may vary based on what’s going on, and I’ve factored this into my quote, which i prefer to define by hours broadcasting.
Some feedback collected from various CoE members (mostly paraphrasing):
It would be more useful to have shorter format how-to’s, example scripts, tutorial articles. This format would be too long for most people to be able to watch all of. Is there a more focused format with with clear takeaways and specific lessons to be learned?
Is this work that would be done regardless. What would the CoE contribution be used for?
I’ve mentioned some clear takeaways, including scripts. A goal I have during this open development is to make an example script for Signals in R. Sure the format isn’t accessible to the most people. It is the format I’m interested in pursuing. Short how-to’s, tutorial articles are a completely different thing.
This gets a big sad face reaction from me. It’s true, I had done this uncompensated in the past, and was planning to do more. I made this proposal because my understanding was that Richard, and some council members, wanted the council to support community-driven content. When I started producing this content, I had hoped to develop a revenue stream for it. The CoE contribution would serve as incentive for me to produce this content and stick to a schedule. If the CoE decides to help fund this project, I’d be glad to take suggestions for what is included in the open-development sessions (and any code or docs developed would be subsequently open-sourced). I don’t know if I’ll do this development live or share the results if I don’t derive income from it. I think there’s a lot to be gained by the CoE making this investment, and hope you all choose to support it.
Just pasting some stuff here from a DM I wrote : "
it’s work that I would definitely do if I were funded. otherwise it would depend on how I feel that day, etc. my goal’s always been to develop compensation
and in the long run, I would definitely not do this project if it weren’t generating income for me"
I don’t disagree, but I see @liz’s work to date as more prototype/proof-of-concept (i.e. a beginning) rather than an end product in itself. I think they did a good job in that respect, and I know I definitely could have benefitted from such when I started.
Evolving from the prototype to the end product—which includes, for example, adapting to the needs of the potential audience through feedback—is quite time consuming, and definitely isn’t the sort of work “that would be done regardless”; otherwise it would have been done by now, no?
Are you withdrawing for personal reasons or lack of feedback from CoE or community? Is it lack of R people support? Not sure if you should withdraw so soon. I think this can be a good thing. Its possible there is people who love to code in R and could get pushed away seeing everything in python possibly too. I think if you took some of the feedback from @hb_scout, a little shorter but possibly more frequent streams taken up in chunks etc. Funding for this should come from CoE for sure. Yes, @hb_scout states that people might be doing what you are doing already but its true R examples is scant. Maybe this would not be a long term thing but more of a shorter term project to expand on github for R people?
@liz Yeah, same questions as @aventurine from me. We’re smoothing out various processes, but we’re still sloooow. I’m very interested in funding content. Obviously depends on why you’d like to withdraw it and you always can, but we’re definitely still churning here.
I don’t think there is anyone doing anything like this. I withdrew because I felt the measure had little support and was unsure if I’d receive feedback if the measure became dead in the water, and the way my brain works, it’s better for me to set something down than leave it up in the air uncertain.
I would still love to do this project, and I like the parameters I’ve proposed. If the council wanted to propose it to me, I would very likely accept, or if they were ready to vote and wanted me to resubmit, I’d gladly do that.
Here’s the deal in my eyes as a person outside the CoE elder circle, if that means anything, but someone who has faith this thing will work. This thing is new. New things take time to move and time work out the kinks. This is perfect excuse for one of the CoE to propose being a “facilitator” of sorts moving these things through. Fact is, and please tell me if I’m wrong, @richai wants to make it “rain” NMR and create positive feedback loops. Proposal in, work done, payout, new people in/Numerai project exposure/tighter community. This is Numerai treasury funding and not coming from anyone’s personal wallet. I think we’re still treading water too much still but things are starting to take shape(kashi paper and prediction marketplace and marketplace web address vote) but, a facilitator, someone to call CoE meetings, create dialogs in the proposals, advancing edits to proposals etc is the way to go. Mechanically speaking, the CoE is only the “final” say anyway. They should Not be the ones who should be starting the initial vote unless it’s one of their own proposals, just dialog and suggestions for edits. We should be 1) making proposals 2) awaiting feedback from CoE members and community 3) make final edits 4) the person who made the proposal start a community vote 5) CoE votes on community passed votes to make final say 6) get paid 7)Numerai gets more outside exposure/ community stays tight.
Things would get filtered out that don’t make sense even before hitting the CoE, but, the CoE is that safeguard in case of proposal attacks etc, eg say someone registered 1000 emails to the forums and started a proposal to create Twitter bot to spam crypto community to buy NMR “NMR Moon Soooon!”, started vote and used those 1000 accounts to vote yes, the CoE can still stop it.
There is no way to know but an interesting thought, how many R programmers have stopped by, browsed site and GitHub and said to themselves “not enough info” and took off. Even something like this proposal once shaped and formed correctly can have that positive loop in my eyes.
Look at my proposal for a newsletter. No action really at all but I think once things kick off the ground even more especially if we had a CoE facilitator it could get some feedback for me to edit and start up a vote
Hi Liz. I’m not a R programmer, but I would love to watch it (live if it falls reasonably within my time zone) if it is short/semi-long-form with a language agnostic topic. E.g. a 1hr stream about methods and things to watch out for when cleaning the data, language agnostic topic but implemented in R.
Strongly agree with @aventurine. This is a proposal from someone who is a well-known member of the community. The requirements for a proposal like this are straightforward, just get it going and provide feedback as it goes. If CoE doesn’t like the direction you can stop paying. Getting community projects active is way more important than getting things perfect at this stage.
This proposal is low risk with lots of upside. Just echoing what @aventurine is saying, but community projects have compounding and network effects. I would love it if we were paying several users to make streams, youtube, and written content. All of that content can build off of each other to reach new users and create new ideas.
Most importantly, Richard has stated how much Signals submissions are helping the fund despite how few Signals submissions we are receiving. Signals growth is crucial for the growth of Numerai. There should be an emphasis on proposals that help with Signals growth. Step 2 of the master plan
