I would also like to point out the (maybe?) in my main post about this. I don’t purport to know with certainty that anything would change regarding these topics but I did want people to be aware. There exists a non-zero probability that I would still be unable to cover certain topics, but the change in sponsorship and independence of the CoE (maybe?) gives more freedom.
1 Like
As you know, DSC and OHwA is so much more than just pontificating about scores, but your point is clear. I don’t ever want the CoE to be calculating the NPV of a person’s value per hour. Then we enter a race to the bottom.
3 Likes
I agree with that and have to admit I was probably coming at this from the wrong (a different) angle. I’m hard wired as a former boot-strapped entrepreneur and money manager for risk assessment and fiduciary responsibility. I’d make a terrible VC and I would suck inside a start-up because I really don’t have the “let’s chase it” mindset. I think and re-think and over think everything until I’ve killed it. That’s completely counter to the internal culture of Numerai. I’ve heard Slyfox more then once express his belief in “rapid development”, we’ve heard Richard express the same thing and now we’ve heard Chanes express the same thing. They want to fund development, a lot of development and I think their willingness to fund the CoE over the long-term may have as much or more to do with how many projects get funded vs how many succeed. With that in mind, I’m changing my vote to yes for 120 NMR a month, but do it knowing it probably doesn’t and shouldn’t matter. We propose, the CoE funds, Numerai will watch and no doubt speak with their wallet.
4 Likes
Anecdote: I always look forward to DSC. Having a way to ask questions live and get them answered has been extremely helpful for me, especially when it comes to signals.
I have a very limited background in financial markets or quant finance. The whole point of numerai was that you don’t need that to join the competition for classic. For signals though, having someone who is knowledgeable about this stuff has helped me a lot.
He’s had lengthy discussions about where to get data, why a longer target horizon might be good (which seems to be coming), creating a signal that’s not a model trained off targets among many other topics. I would encourage anyone who is interested in signals to listen to past DSCs. You will learn a lot.
3 Likes
I think this is a very strong point. Current balance of CoE wallet is 58 NMR. This means CoE isn’t even able to fund this proposal for a single month in its current state.
I understand this falls a bit outside the scope of this particular proposal, but where is the next CoE’s funding coming from, at what frequency and on what terms? As long as we don’t know this how can we discuss allocation of budget?
3 Likes
I believe it’s coming directly from Numerai. In the CoE video, I think I heard Richard say $100k worth. At this point, we need to get proposals into and through the CoE. At that point it’s up to Numerai to pull the trigger on a transfer that will support the proposals. Our job now, I think, is to get the proposal stage ramped up. The CoE can’t engage in rapid development if no one is dropping anything in their laps.
1 Like
we’re quite off topic, but the Council can also post “request for proposals” and allow for bids and we’re not set up for that just yet. This is something I want to see implemented very soon.
1 Like
Wen CoE vote!? I also got to say that as someone without a data science background, the daily scores and chill really helped me refine my modeling and continues to this day while I continue to work on making a signals model. When I would hear talk about data science topics that ive never heard of before and heard them explained over steam, it gave me that extra motivation to look into the topics further. I remember specifically a stream last year that arbitrage and another member were on talking about feature selection and using supervised learning modeling with scikit by mixing together linear and non linear models using a pipeline. That specific one helped me a great deal with one of my models. The daily scores is not all about scores. The office hours with special guests is really cool too. Gives you an idea of what others are doing and it lets us all get to know each other a little more. If breaking away from Numerai is needed for this to become more open and being funded by the CoE, im all for it! Vote, Vote, Vote!
1 Like
Hi Jon,
Is this an all-or-nothing proposition? Or is the price negotiable? I can definitely see the value in this but 120 NRM per month seems a bit much, and may deter other new proposals if most of the CoE budget has already been used up.
Regarding the KPIs you’ve included, are you proposing that you get paid the full amount regardless if the proposal was deemed a success or not? It might be worth considering having part of the amount as ‘base’ and the rest as a type of bonus incentive structure, depending on if it was deemed a success or not (by the CoE or the Community I suppose)?
2 Likes
Also, even though you are not voting yourself, it may have at least the optics of a conflict of interest.
1 Like
I have enjoyed and benefited from the ohwa and dsc streams and would really hope they would continue. My one suggestion for any content project is for the CoE to greenlight them one season at a time (however CoE defines a season).
5 Likes
I voted No to CoE being the sponsor. This is why: trailblaze-vs-scale. My belief is that CoE should be in the business of solving the lack-of-enough-disruptive-ideas problem, by seeding many new and crazy pilot projects.
I think OHwA and DSC are valuable, delightful, should definitely continue, and it is Numerai’s job to resource it. Just like it is Numerai’s job to scale-up the best ideas that graduate from the CoE.hatchery.
2 Likes
I am leaning towards this type of philosophy as well. CoE shouldn’t just be taking over functions that are naturally Numerai’s. But that’s in general. I could go either way on this one – at least for OHwA. DSC is fun but it is really kinda limited to a small subset of people in the right time zone (and the right day jobs). Do many watch those streams after the fact if they can’t do it live? OHwA is more of a resource (although they too get outdated after a while, but not right away).
If it’s about being compensated for the workload, I’m wondering if we could decentralise Office Hours and DSC, to share the load (and pleasure) of hosting. Perhaps with a rotating chair and region, starting with different members of the CoE? Having built in redundancies is generally a good thing to avoid single point of failures. Though maybe there really is more to it than just hosting a zoom call and hit record?
/Joakim
While I’m sympathetic to the general feeling that paying Arbitrage USD 3,000 per month in NMR seems like a lot - that’s a real salary for many (I have read in posts above that it’s estimated at USD 46 per hour), we have to keep in mind that he’s not an hourly wage worker. He’s shown that he’s highly skilled at producing these videos and creating engagement within the community.
I’m quite certain that investors in Numer.ai today would argue that the engagement and exposure created by the long duration content of office hours (it’s not really evergreen but I’ve gone back to see 2 year old videos. Aven’t you all?) produced by Arbitrage far exceeds USD 36,000 per year over the last 2 years. Therefore paying him the same amount to continue going forward seems like a no brainer to me. I think anyone that holds NMR is indirectly benefitting from his efforts.
3 Likes
Maybe you’re right.
However, for me, it’s not so much the absolute amount I have a problem with. I completely agree that Arbitrage has created > $36K of value per year. No doubt, 100%. I also agree that he deserves to get reasonably compensated for all his efforts (I’ve benefited as well).
It’s the relative piece of the pie that I think is disproportional. If the CoE budget was NMR 20,000 per year then yeah, it’s a no-brainer to me as well. Now I don’t know what the CoE yearly budget is (no-one seems to know this), but if it’s NMR 2000 / year, then paying a single person NMR 1440 (72%) seems a bit much. What about @uuazed who created Numerapi? Or @surajp and @jrai who created amazing Signals example scripts? Or @ceunen who created the amazing Numerai Payouts app and dashboard? I could go on. Aren’t they equally deserving of a decent piece of the pie? Maybe they got paid directly by Numerai, in which case transparancy and communication around this could definitely be improved.
In short, I think the CoE needs to first work out what their mandate, budget, KPIs, etc. are before approving large (relative to budget) payouts to a single person.
2 Likes
Yes. I understand and agree with you Joachim.
The issue is more about what is the financing mandate of the COE vs Numer.ai itself, and what other projects the COE could finance instead with the same resources.
2 Likes
I am trying not to be offended or to take anything personally, because I know no one is making attacks. I struggle with defending my compensation, because that’s not a good conversation in any context. I don’t like where this conversation is going and I simply ask that you refrain from converting human beings into commodities.
Whether Numerai retains my contract or the CoE picks it up, the outcome is the same. Regardless of which entity provides funding, I will continue to provide the highest quality content that I can. Content which encourages exploration, ideation, and creativity. Content that teaches. Content that is timely. Content that fosters collaboration. Content that entertains. Content that builds community.
I can’t vote on this proposal and I can’t place the proposal in the queue to be voted on, either. This proposal is dead until someone who isn’t me (SWIM) on the CoE moves it forward. Reading through Rocket Chat and the forums, I would be surprised to see anyone on the council move anything forward.
2 Likes
I think one of the general points is important: if the CoE simply takes over functions that Numerai the company ought to be doing or would be doing anyway, then we have gained nothing by having a CoE.
4 Likes
Indeed, if opacity is desirable, then let’s just disband the CoE and have Numerai hire someone to fund projects that Numerai deems desirable for the community.