Decide the direction of TAI R&D


Choose the first top priority for TAI R&D.

In a hurry? Don’t read what’s below. Just vote Yield to increase capital efficiency, vote Governance to become trustless and permissionless, vote Stability to for a more predictable TAI price, vote Integration to build bridges with other protocols.

They all have roughly similar labor costs, so you can vote entirely based on the expected benefits.

1) Yield Surge

Without increasing risk, increase capital efficiency & yield. There are likely some tiny improvements available from tuning parameters. Major advancement over existing protocols requires innovation. Before joining TAI this was an issue I was already working on, and I have 2 promising methods.

The first negligibly increases risk while increasing TVL and yield by an estimated 5-15%. This first method has low implementation cost; it can be be assembled only using existing audited solidity. While the code is simple, it requires a lot of planning.

The second innovation method is to consider the perpetual swap to be the theoretical maximum efficiency. The apparent issue with perpetual swaps is that they have notional value dozens of times their collateral. While this risk is largely removed by requiring equal notional value long and short, perp swaps are still able to black swan. Consider however that the large increase in yield from being many times more capital efficient can be spent undoing the risk. When I deeply studied the largest perp swaps during the covid market crash, their insurance funds only needed to be 0.11% of the total notional long+short value. I am thinking it is plausible to double yield and still have enough left over to insure a black swan chance to once per 10 thousand years. Perhaps this can be built with just parameter changes of our current system and a couple simple module addons, but we may find it requires a labor intensive total redesign of perp swap solidity from another project.

If we vote Yield Surge, almost all effort will go into the first method. It is within grasp. The whitepaper is partially completed. We can copy paste parts of our existing code to make it. If new code is needed, it will be less than a screen full.

2) Governance Purge

Efforts to minimize or eliminate governance. At the center of this is risk assessment and approval for new collaterals. Under the highest standards, full protocol autonomy is required to considered a DAO. If humans can interfere with the protocols mission in any way including by taking over governance voting, it is not autonomous.

This could bring in a surprising amount of investment from OG types who do not take any protocol seriously unless it is fully trustless and permissionless.

3) Verge Stability

Research into the profitability of arbitrage in order to intentionally configure the system to have healthy competition between arbitrage. Stability comes at a cost; a rent paid to arbitrage. This R&D might result in a new mechanism used by arbitrage. For example one that makes their decisions less ‘statistical’. It could also simply result in improved parameters for the interest rate controller.

A stable price is generally required for adoption for use in high leverage and commerce. Stability will increase demand for TAI; however the benefits take years to manifest.

4) Integration Splurge

Research into how our protocol can benefit other protocols, and clever ways we might utilize other protocols for collateral or value to our users. This research would include questioning the ideal ways to give out our subsidies to LPs. One possible outcome of the research is a small protocol change to enable compatibility, but most likely this will simply be searching for the best protocols to cooperate with, coding UI or simple backends for ease of use. Then if liquidity is critical, a bit of marketing effort or subsidies to get the ball rolling.


  • Yield Surge
  • Governance Purge
  • Verge Stability
  • Integration Splurge
  • There are better suggestions in the comments below. Remake this poll so we can vote on those too.
0 voters
1 Like