search

Tribler / tribler

Privacy enhanced BitTorrent client with P2P content discovery
https://www.tribler.org
GNU General Public License v3.0
4.74k stars 445 forks source link

literature survey + master thesis: G-Rank learn-to-rank #5313

Closed synctext closed 1 year ago

synctext commented 4 years ago

Direction changed, txt will be updated soon.

Old stuff:

awrgold commented 4 years ago

tragedy2.pdf

devos50 commented 4 years ago

Note that the tragedy of the commons is even present in blockchain systems (as a consequence of full state replication).

awrgold commented 4 years ago

New Literature Survey:

Autonomous (decentralized?) trading bot:

synctext commented 3 years ago

True autonomous AI has full freedom to act, own and profit. This literature survey is focused on ...

Links

awrgold commented 3 years ago

Next step:

~10 sentences or so, narrowing the scope down to 3-4 directions to go in regarding the survey and prototype. (Next week)

synctext commented 3 years ago

This is the recent breakthrough in reinforcement learning algorithms (e.g. Asynchronous Proximal Policy Optimization (APPO) algorithm). Entire thesis goal/Survey-prototype could be to get this into the superapp AI bot, get a Tweakers post, deployment to a few hundred phones, measure, plot in thesis, and should scale to millions. Each smartphone has fixed 5-ish GByte slice of a huge dataset of historical market data, making each bot unique. Creates a very concrete focus: apply this ML technique to autonomous trading bots. Specifically, smarter bot trading this recent work on our decentral market; see here. Fancy results. Done. (only fundamentally re-work this technique by distributing the shared memory and decentralising the learner to offer unbounded scalability. Simple. /sarcasm)

Key research question: Can we map the “Sample Factory” (a high-throughput training system optimized for a single-machine setting) architecture to the Autonomous trading bot setting?

image Website: https://sites.google.com/view/sample-factory Article: https://arxiv.org/pdf/2006.11751 Press: https://spectrum.ieee.org/tech-talk/artificial-intelligence/machine-learning/powerful-ai-can-now-be-trained-on-a-single-computer Code: https://github.com/alex-petrenko/sample-factory Demo:

awrgold commented 3 years ago

Hey Johan, This looks interesting at first glance. I will start reading up on this before our meeting tomorrow. I just got back from a super spontaneous short holiday with some friends, so the last 5 days I have been away. I did have some great discussions with my friends though on this vacation (we all did our data science Bachelor together) so I am thinking I'm narrowing down the focus. Reinforcement learning is obviously the Mecca of "proper" AI as the theoretical ceiling on its potential is astronomical compared to (un)supervised learning, but of course is far more complex than simpler techniques. I will share more thoughts tomorrow.

Current focus: adversarial training for trading bots, GANs, generative modelling, autonomous/online learning, q-policies, independent modelling, etc.

See you then, Andrew

On Sat, Jul 18, 2020 at 11:03 AM Johan Pouwelse notifications@github.com wrote:

This https://arxiv.org/pdf/2006.11751 is the recent breakthrough in reinforcement learning algorithms (e.g. Asynchronous Proximal Policy Optimization (APPO) algorithm). Entire thesis goal/Survey-prototype could be to get this into the superapp AI bot https://github.com/Tribler/trustchain-superapp/blob/master/README.md#ai-trading-bot, get a Tweakers post, deployment to a few hundred phones, measure, plot in thesis, and should scale to millions. Each smartphone has fixed 5-ish GByte slice of a huge dataset of historical market data, making each bot unique. Creates a very concrete focus: apply this ML technique to autonomous trading bots. Fancy results. Done. (only fundamentally re-work this technique by distributing the shared memory and decentralising the learner to offer unbounded scalability. Simple. /sarcasm)

Key research question: Can we map the “Sample Factory” (a high-throughput training system optimized for a single-machine setting) architecture to the Autonomous trading bot setting?

[image: image] https://user-images.githubusercontent.com/325224/87848851-1c3bc800-c8e4-11ea-8954-fe7a3521a423.png Website: https://sites.google.com/view/sample-factory Article: https://arxiv.org/pdf/2006.11751 Press: https://spectrum.ieee.org/tech-talk/artificial-intelligence/machine-learning/powerful-ai-can-now-be-trained-on-a-single-computer Code: https://github.com/alex-petrenko/sample-factory Demo:

https://github.com/alex-petrenko/sample-factory/blob/master/gifs/battle.gif?raw=true https://github.com/alex-petrenko/sample-factory/blob/master/gifs/duel.gif?raw=true

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Tribler/tribler/issues/5313#issuecomment-660453135, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADFXMZWJLFO4ZUJE3FUUZ4LR4FQNDANCNFSM4M3KKCZQ .

synctext commented 3 years ago

As a first survey step: please post here for Monday 24Aug: .PDF file in IEEE 2-column format with 20 citations of papers and first draft text in sections. Plus 1 giant table with all papers, per year, expanding something like: image Ideally add 1 graph of performance of experimental part. Idea: make a bandwidth trading bot ecosystem emulator with self-play. So 10 bots which do stuff and can be used to compare algorithmic complexity, intelligence + resilience against attack.

synctext commented 3 years ago

Can you find more in this: https://papers.nips.cc/paper/5656-hidden-technical-debt-in-machine-learning-systems.pdf ? "Hardest part of AI is not AI". You agree?

synctext commented 3 years ago

5 x 5ects courses left. plus lit survey. Proposed direction:

awrgold commented 3 years ago

Interested in quantification of activities to help inform existing financial/investment funds regarding climate change, ecological destruction, financial regulation, Pigouvian taxes and climate regulation, etc. - "making the world a better place through data?"

"What would a Climate DAO look like?"

Papers and literature to follow.

Long-term TODO:

synctext commented 3 years ago
synctext commented 3 years ago
synctext commented 3 years ago

@awrgold : 'i like making trading strategies' liquidity pools, yield farming, :floppy_disk: Tribler Credit Farming :+1: Liquidity pool: IBAN Euro vs. Bitcoin?

synctext commented 3 years ago

ToDo: superapp compile from the source. Understand Euro Token, liquidity pool, Taproot, and keep track of documentation/grey_literature. With bit more background info and status understanding we can define the exact topic in 3-4 sentences.

awrgold commented 3 years ago

I hope you don't mind me using this issue as a personal bulletin board for some thoughts. Q4 just started, doing some reading today to get back up to speed. The last 12 months were a blur, I'm trying to get back not just on track but also making sure that what topics I do chose will keep me motivated for a longer period of time.

I'm (re)reading a lot of the Tribler/Trustchain documentation and development history, catching up on what's already been attempted, completed, shelved, and abandoned in the past. Particularly this page. As such, here's some topics I think are either useful and/or interesting.

Topics of particular interest:

synctext commented 3 years ago

Discussion of thesis direction

awrgold commented 3 years ago

Ultimate goal: graduate by end of Q3 2022

synctext commented 3 years ago

Final possible deliverable: report co-authored with Authority Financial Markets...

synctext commented 3 years ago

Week 5 of this quarter currently. Making progress with investigation of current DAO work. Goal of thesis and survey concrete (repeating): production-ready DAO with 1+ Bitcoin, 100+ people and BIP docs of your Taproot+trustchain coordination protocol. Focus on getting the superapp code from 'class project' to financial critical infrastructure level. What does that actually mean and how do you show that you delivered that? (Tools? https://nexocode.com/blog/posts/mutation-testing/)

ToDo:

awrgold commented 3 years ago

DAO Literature

  1. MakerDAO Whitepaper: https://makerdao.com/en/whitepaper
  2. List of Active Ethereum DAO's: https://app.daohaus.club/explore
  3. The DAO and Governance: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3082055
  4. Bitcoin and the rise of DAOs: https://link.springer.com/content/pdf/10.1186/s41469-018-0038-1.pdf
  5. Experiments in Governance: https://moodle.epfl.ch/pluginfile.php/2861870/mod_resource/content/1/DUPONT-2017-Preprint-Algorithmic-Governance.pdf
  6. Smart Contract Collaborations: https://link.springer.com/chapter/10.1007/978-3-319-21915-8_1
  7. E-Gov DAO: https://www.researchgate.net/publication/325632774_eGov-DAO_a_Better_Government_using_Blockchain_based_Decentralized_Autonomous_Organization
  8. fetch.ai: https://fetch.ai/

Infrastructure Literature

  1. ZK-rollups: https://docs.ethhub.io/ethereum-roadmap/layer-2-scaling/zk-rollups/
  2. MEV Crisis: https://medium.com/flashbots/frontrunning-the-mev-crisis-40629a613752
    1. https://ethresear.ch/t/flashbots-frontrunning-the-mev-crisis/8251
    2. https://medium.com/etherscan-blog/rapid-rise-of-mev-in-ethereum-9bcb62e53517
  3. MEV: https://blog.chain.link/what-is-miner-extractable-value-mev/#:~:text=One%20such%20example%20is%20Miner,excluding%20transactions%20within%20a%20block.
  4. Aragon: https://aragon.org/protocol
  5. DAOstack: https://daostack-1.gitbook.io/v1/

Other Literature

  1. Private Transactions on ETH: https://tornado-cash.medium.com/introducing-private-transactions-on-ethereum-now-42ee915babe0
  2. Ethereum 2.0 in Rust: https://github.com/sigp/lighthouse
  3. COPE Tokenomics: https://www.unlimitedcope.com/assets/COPE_Tokenomics_0.pdf

This is what I've compiled so far on DAOs, I have like 50 other papers regarding blockchain and cryptography and the like, but were more focused on the earlier iteration of the literature survey from last year, e.g. trading bots, etc.

synctext commented 3 years ago

Production-DAO needs a service. Smart contracts need to scale somehow. Possible idea is server renting and compute servers in an open market using Bitcoin. See prior working code and the [Plebnet report]. (alternative is a simple btc/eth trading bot). Scientific interest focus of thesis: governance?!? The self-governance of this service is designed to be a role model for economic activity under direct democratic control. ToDo: scoping for 1st phase? Keep exclusively DAO for literature in narrow scope (determine historical development timeline, investment capital tracking), no NAW, decentral economy fetch.ai out of scope?
Papers found: https://doi.org/10.1109/SDS49854.2020.9143877, https://doi.org/10.1109/ACCESS.2021.3081926, https://doi.org/10.1787/ef4eba4c-en, https://doi.org/10.1109/TTS.2021.3054974 (Measuring DAO autonomy, solid overview paper) . idea Copy and expand upon using performance evaluation: https://arxiv.org/abs/2011.14940 image Example of a format that I have in mind from another students; survey, [64] citations, screenshots, experimental evaluation, thesis pre-prototyping https://arxiv.org/pdf/1512.00071.pdf

synctext commented 2 years ago

Possible methodology: don't build fantasia about how a DAO could look like. Others are doing that for us :laughing:. Do a hands-on production-DAO without corner-cutting. Bottom-up and driven by a single actual service. Minimal Viable DAO, must have service in the machine economy of the future. Server infrastructure (e.g. Bitcoin VPS). Reversed cloud: not owned by corporation, not owned by humans, but DAO controlled. (https://akash.network/). Other must-have alternatives are a global key-value store: https://blog.sia.tech/skydb-a-mutable-database-for-the-decentralized-web-7170beeaa985 Dependency, that requires servers. Is the taxation status in scope? Keep strictly technical (distraction: ask for a formal tax ruling on Delft DAO?) Tax distractions:

Happening: https://www.coindesk.com/yield-guild-games-dao-funding-round-delphi-scalar

synctext commented 2 years ago

Blind brainstorm on master thesis focus. Working backward from Math&ML talent and thesis outcome. Literature survey will obviously be broader. Required background reading: Delft DAO Taproot implementation

The 2002 Nobel Prize in Economics was awarded to Vernon Smith, an experimental economist. This award represented
a turning point for the visibility of economics as grounded in experiments. Computer science is also transitioning,
Big Tech platforms now dictate the laws of economics. Trillion dollar companies now define the rules, regulations,
and economic principles within their ecosystems. This thesis will create a non-profit platform which, in principle,
could grow beyond a trillion dollar.

Decentralized Autonomous Organizations represent the future of automation in the upcoming robot economy.
This master thesis will use the experimental economists methodology to explore the future of how our economy
may be organised, who will own it, and if taxes will be paid. Starting point is the Internet-deployed experimental
Delft DAO. The primitive self-governance of Delft DAO is designed to be a role model for economic activity
under direct democratic control. A early operational prototype of Bitoin Taproot is completed in Delft DAO. 
This thesis will first create a design and matching byte-accurate mathematical model of Delft DAO multi-sig
transaction size, cost, and scalability. The design will build upon Bitcoin Taproot and Trustchain. Second,
implement this model. Third, devise machine learning techniques to improve one aspect of Delft DAO, such as,
robustness, scalability, fraud-resilience, or attack-resilience in general.
awrgold commented 2 years ago

https://en.wikipedia.org/wiki/Unified_Payments_Interface#Acceptance

Saving for later

synctext commented 2 years ago

Survey of "DAO" for 10 ECTS:

awrgold commented 2 years ago

Roughly where I stand regarding (online) sources. I have the names of some published papers in the .md file, but I didn't want to upload those to Github and I still have many more sources covering (often tangentially) relevant topics for DAOs, governance, infrastructure, economics, etc.

https://github.com/awrgold/DAO_literature_survey/blob/main/Literature%20Survey%20Sources.md

synctext commented 2 years ago
Example of output? Pick the Top-3. Wireshark the real technical stuff from your app/browser/wallet. Key to explore centrality and fragility. DAO name ERC-1155 Scam Alert Level Founders Rich First Screenshot Critical Central servers Just a cloud app, nothing decentral
MakerDAO :x: :heavy_check_mark:
traderjoexyz . com :x: :heavy_check_mark:
Uniswap1UNI :x: :heavy_check_mark:
MetaCartel :x: :heavy_check_mark: :x: :heavy_check_mark: :x: :heavy_check_mark:

Next sprint: no writing please :memo: , screenshot joining, screenshots voting, hot wallet addresses, IPFS central pinning, do a vote with Wireshark, scan voting history for anomalies, identify central failure points. Select your final top-3 for deep-diving.

awrgold commented 2 years ago

PunksComic, MetaHero, and PixelVault are working together to raise (from the community) approximately $200m to develop an NFT-based AAA-quality game based on the #1 NFT project in the world, CryptoPunks.

https://punkscomic.com/planetdao.html

Those who own punks (the cheapest being around $250,000) will have exclusive privileges in the game, but the community crowd sale is of new NFTs for their video game who then can use these as exclusive token-based gated privileges into their community.

You can fractionalize Punk ownership, however: https://decrypt.co/68604/50-cryptopunks-nfts-go-fractional-in-new-collection

awrgold commented 2 years ago

IEEE format paper on Overleaf: (empty right now) https://www.overleaf.com/read/vdftyrvpjzyc

awrgold commented 2 years ago

TODO:

awrgold commented 2 years ago

I didn't even realize it, just stumbled across this. There's a live remote Ethereum summit regarding DAOs, particularly discussing legal/regulatory environments + governance structures. https://www.youtube.com/watch?v=szg_CvrxU9w

(it's live right now as of this posting)

awrgold commented 2 years ago

So joining a DAO is proving to be somewhat challenging, as I often need to:

  1. Purchase tokens on an obscure DEX and stake them in a pool
  2. Submit a proposal for membership
  3. Doxx myself using something like https://www.brightid.org/

Now I'm trying to find DAOs that allow anonymous instant public membership.

synctext commented 2 years ago

3. Doxx myself

:dagger: :dagger: :dagger: Watch out for that please. 99% of all projects are either scams, naive utopia builders or both.

awrgold commented 2 years ago

This is true, looking more into BrightID it seems to be a tool that I can use to verify my identity through what they call "verification parties" but without having joined one yet, it seems like it's just a Zoom call to show that you're a real person. Doesn't seem too decentralization-friendly, in my opinion. I'll keep looking.

The tokens thing isn't prohibitively expensive, it's the ETH gas required to make it happen. When I did an estimate for bridging ETH -> DAI -> xDAI, it would cost about $225 in gas. Gas prices are insane right now (both irl and in the metaverse) because of all these NFT mints and increased DEX volume, so even simple swaps via a DEX can run upwards of $100 in fees.

awrgold commented 2 years ago

Okay, so I'm hooked on these two DAOs called Olympus (OHM) and Wonderland (TIME). The concept flips liquidity mining on its head, creates incentives to continue to provide liquidity to the same protocol instead of finding ways to create derivatives/tokenizations/collateralizations of your assets to start "looping" and farming yield.

HOWEVER, the dev teams are completely anonymous, the whole thing smells like an elaborate ponzi scheme, and the treasury and DAO is controlled by a K of N scheme locked into a time-locked Gnosis safe. As long as that safe is closed, the devs can't touch any of the funds. But during the window that safe opens, the devs can technically withdraw all the liquidity (every OHM/TIME token is backed by a basket of assets including stablecoins and ETH) and just vanish.

The public at large is drawn to the enormous yield provided to stakers. OHM has 7300% APY, rebases every 8 hours, and automatically compounds the yield into your staked assets, so you can "set and forget" and watch your stack grow. TIME is essentially an OHM fork (a copy/paste job) on the Avalanche chain, and provides close to 50,000% APY (that's 2% per day) to stakers. (At one point earlier in June, OHM had more than 200,000% APY - 8% interest delivered PER DAY).

As of right now, I'm staked (an affordable amount) in both and have timers set for before that time-lock contract unlocks, so I can withdraw before the devs get access to the private key.

The community is noticeably (overly) bullish - I mean, who wouldn't be with that kind of yield - but detractors and genuine criticism is going unanswered in the community, besides for vague promises of "the K of N scheme being addressed publicly Soon:tm:" so I have a pretty good feeling this is a scam.

The thing is, between these two schemes there's over $500m in ETH/stablecoins locked into these contracts. If the devs pull the rug, they could easily pull off one of the largest crypto heists ever performed. Of course, all of these addresses are public and the tokens/addresses can be blacklisted, but yeah it's kinda like watching a train wreck in slow motion - a train wreck someone can make insane profits with by watching from afar and walking away right before impact.

OlympusDAO audit #1: https://omniscia.io/olympusdao-algorithmic-currency-protocol/ OlympusDAO audit #2: https://github.com/peckshield/publications/blob/master/audit_reports/PeckShield-Audit-Report-OlympusDAO-v1.0.pdf

OlympusDAO dashboard: https://app.olympusdao.finance/#/dashboard Wonderland dashboard: https://app.wonderland.money/#/dashboard

synctext commented 2 years ago

Epic story! Please dive deep and measure everything, trace it all please. Bundle in 6-8 IEEE 2-column pages. https://cryptosec.info/exit-scams/

History repeating: "Fraudsters abandoned the scheme by withdrawing over $3 Billion dollars in Cryptocurrencies (Bitcoin, Ethereum, and EOS) and leaving the message “sorry we have run“."

Great prior taxonomy work https://core.ac.uk/download/pdf/301379414.pdf

### Taxonomy of crypto scams Within this literature survey we clone a DAO clone and successfully conducted an exit scam. We conclude that no technical measure is keeping the DAO assests safe. Security critically relies on the DAO developers being uncorruptible saints and developers with zero-defects. The latest generation of DAO technology operates outside the reach of the law. Irrefutable proof that they have real Bitcoin and Ethereum? Wonderland (TIME) DAO has 500 Million $ worth of Bitcoin which critically depends on anonymous strangers who may exit scam at any time without jailttime consequences. | DAO name | clone without additional value | Scam Alert Level | Founders Rich First | Screenshot | Critical Central servers | Just a cloud app, nothing decentral | self-delusional utopia | | --- | --- | --- | --- | --- | --- | --- |---| MakerDAO | :x: | :heavy_check_mark: | | | traderjoexyz . com | :x: | :heavy_check_mark: | | | Uniswap1UNI | :x: | :heavy_check_mark: | | | MetaCartel | :x: | :heavy_check_mark: | :x: | :heavy_check_mark: | | :x: | :heavy_check_mark: | 2 week focus: a 1 scam story plus scam table {25 citations}; be 50% done with writing; overleaf pages; block-scanner work :1st_place_medal:
synctext commented 2 years ago

Feel free to mention that a DAO with tax compliance would really alter the world. $0 corporate income tax for Amazon. An estimated US$21 to $32 trillion of private financial wealth is located, untaxed or lightly taxed, in secrecy jurisdictions around the world. Only thing needed is obsolete central bank based money and transition the global financial system to DAO-based economy with global democratic governance :boom: :medal_sports: :boom:.

awrgold commented 2 years ago

I am in the process of scheduling a private meeting with the OlympusDAO founders. I told them that I'm diving deep into their project and trying to understand their economic model, and that I'm a grad student working on a thesis in financial DAOs with a potentially large audience, and they reached out to me directly and said I can ask them direct questions anytime I want, and if I had a list of questions we could schedule a meeting so they could answer them directly.

I'm now working on trying to finalize a checklist of questions I have, particularly revolving around the security of the DAO/keys, as well as the tokenomics involved.

awrgold commented 2 years ago

image

Someone who had over $20m in OHM tokens started selling, causing everyone to freak out and sell as well, "oh no the ponzi is ending" and well, the price collapsed -50% in about 4 hours.

synctext commented 2 years ago

LOL, buy into this scam cheap!

awrgold commented 2 years ago

https://olympusdao.medium.com/introducing-olympus-v2-c4ade14e9fe

"Olympus v2" was just announced with major changes announced to the protocol, bonding mechanism, vesting schedules, rewards, DAO governance, and more. One of the primary changes is that there will be a token migration. Apparently one or a series of whales decided this was a good time to take profit and market sold on SushiSwap, creating a cascading effect.

Additionally, many people are leveraging their stakes by "looping" via borrowing against their staked assets and then staking their borrowed claim. I haven't done any on-chain digging into this yet, but it could be that at least some of the price drop is due to cascading liquidations (remember there's no margin calls in Defi).

That being said, I'm reading and discussing these changes proposed in v2 with those in the Discord to try and understand what new dynamics they will introduce. As time goes on, the more I believe that this is not an intentional scam, and instead just a protocol with certain positive feedback loop features that can be exploited for profit.

Primarily, the Olympus meme (3,3) refers to the prisoner's dilemma, and many anons have (3,3) in their names or bios. This refers to the idea that if you (X) and everyone else (Y) both stake your OHM and don't sell, it will benefit everyone else. However, this of course is fallacious as in the prisoner's dilemma the highest EV comes from ratting out your accomplice (in the case of OHM, selling). However, due to OHM's linear bond vesting schedule (which will be done away with in v2, apparently), a certain kind of optimal frontier begins to materialize around staking for some time until one decides that a local top is forming in the market price of OHM, and then selling in the hopes of lower prices in the future. Whether or not this mechanism will lead to more volatility over time, or become less profitable as the risk-free value (RFV) of the treasury increases, I'm not quite sure yet.

Either way, what we can see is that there was a local top in the market price of OHM around $1350, and after a few large selloffs on 13 October, the announcement of the switch to v2 + the token migration plan may have been enough catalyst to convince an early supporter/whale to dump their tokens. Additionally, as the price collapsed, the price of bonds (which are usually sold at a discount to the market rate, based on demand) skyrocketed to a roughly +15% premium over the market value of OHM, meaning that when the price collapsed, enormous influx of buyers showed up seeking to purchase bonds until all available bonds priced at a discount were eaten up. The premium on the bonds reaching +15% showed that buyers, for at least a short period of time, expected the price of OHM to recover more than +15% of their principal investment. Since the crash, the price of OHM recovered by +30% off the lows, and is currently trying to find a new support.

Bonds have since recovered from the premium and are now selling at somewhat larger discounts than normal. According to many I've spoken with, this is expected behavior and is intended to create some kind of equilibrium over time.

image

Edit: By the way, it's worth noting that as the price collapsed, the yield (APY) jumped by about 15%. I staked (stoke? stuck?) 1.0 OHM token back when it was $600 a few weeks ago. Since then, my stake has grown by approximately 7%. At this rate, I will have doubled my OHM tokens in 57 days. The increase in my holdings somewhat offsets any downtrend in price. As the price continues to decline, the yield increases. Remember: for every OHM token there is approximately $140 in backing, so OHM can never go under $140 (major asterisk here: that dollar value is largely in DAI, another algorithmic stablecoin. If that collapses, OHM collapses) as the protocol will begin spending its treasury to purchase OHM tokens trading underneath the price of its current RFV, and then will burn the token.

awrgold commented 2 years ago

There's another DAO I just found that is another soft fork of Olympus: KLIMA DAO. It's the exact same idea, except instead of tokens being backed by a basket of cryptoassets (as Olympus does) each token is backed by a carbon credit of 1 TCT (Tokenized Carbon Tonnes - read this link). This sounds very "pie-in-the-sky" but is an interesting concept. The number of DAOs popping up copying Olympus' "Proof of Bonding" mechanism is astounding, which tells me that not only is there enormous profits to be made, but many people are inspired by the mechanism and will be iterating on top of it for some time.

@synctext I'd like to propose that the literature survey outline look something like this:

  1. Introduction
  2. Finance DAOs, LPs, and Defi (background)
  3. Novel Finance DAOs and algorithmic currencies
    • MakerDAO
    • OlympusDAO (and its direct copy forks such as TIME)
    • KLIMA
    • Any others
  4. A section on potential directions these may go, and potential up/downsides
  5. Conclusion

Thoughts on this structure? The meat would be on section 3, discussing everything from governance to tokenomics, built on the foundation that section 2 describes. I think section 4 is the one I'm least sure about. After describing the landscape and what is happening, what should be said before offering a conclusion?

awrgold commented 2 years ago

@synctext https://ethresear.ch/t/mev-auction-auctioning-transaction-ordering-rights-as-a-solution-to-miner-extractable-value/6788 https://medium.com/flashbots/frontrunning-the-mev-crisis-40629a613752 https://www.twitch.tv/videos/1180411852

awrgold commented 2 years ago

Alright, so just as a quick summary (since I'm still paring down my long writeup into something that can fit in 8-10 pages), essentially the Olympus Protocol has 3 major weaknesses, and one "minor" weakness that if exploited means the entire crypto universe collapses anyways, so no point worrying about (4):

  1. The majority of the treasury's 7 private keys are owned by the protocol developers for the time being. This is known by the entire community, and is apparently being changed in their upcoming "v2" update. Right now, the DAO owns the minority of the private keys, but the developers can still collude to drain the treasury, currently valued at about $612 million. Apparently, v2 will transfer entire ownership of the treasury to a new contract with 7 new keys all belonging to elected members of the DAO. Details on this are yet to be announced.
  2. A bug/exploit in the smart contracts could lead to a drain of the funds, though this is the case with every ERC contract.
  3. The DAO suffers a coordinated attack (in collusion with 1+ of the developers in v1, or with 4/7 DAO keyholders in v2) in which the treasury assets are drained.
  4. The value of the stablecoins (almost entirely DAI, which is 150% backed by ETH) drops precipitously.

However, as for the tokenomics of the project, there is (from my understanding) a pretty rock-solid scheme that can withstand >99% drop in price such that those who remain staked and don't panic-sell will not only not lose their initial investment, they will walk away with a profit - even if everyone else sells (assuming there is no hack and the contracts work as intended). If 99% of the stakers leave the pool and sell their tokens, the remaining 1% will still own the liquidity of the protocol and the rewards for each individual will skyrocket.

It's a pretty interesting idea, and the reason why the yield is so high is because the treasury essentially pays you such high rewards to stake as incentive to store your tokens instead of actively trade them. The mechanics of the protocol don't break if the majority of people unstake and sell/trade their tokens, actually, as the amount awarded to stakers is chosen by the community and hard-coded into the contracts, so until a new OIP (Olympus Improvement Proposal) comes along, the code will operate as it currently does. The next major change to APY rewards is expected to bring the % down from approximately 8000% down to about 1000%. The last OIP to reduce rewards to extend the runway brought yields down from 15000% down to 8000%.

The complexity of the broader ecosystem is astounding, where the foundation of any one ecosystem is based on top of series of stacked dubious assumptions.

Is OlympusDAO a house of cards? Inexperienced developers tend to make broad assumptions of the underlying code and lack the fundamental understanding of their environment in which they operate to make decisions without creating significant negative externalities that propagate throughout the network.

The yield rate over time looks like this: image

synctext commented 2 years ago

Students which are completely lost, but are forced to hand-in code tend to make bloated designs and bloated code.

awrgold commented 2 years ago

https://docs.olympusdao.finance/main/

synctext commented 2 years ago

Yes, good storyline proposal:

synctext commented 2 years ago

https://adjacentpossible.substack.com/p/lets-run-the-experiment-a-conversation

Now let me address your question about DAOs. The word DAO is probably overloaded
at this point, but in short refers to a new kind of organization that uses blockchains in the
way I alluded to above. In some loose sense open source software projects ranging from
Linux to Ethereum are DAOs. They make decisions about software and network upgrades
via roaming groups of users and developers on various websites discussing and arguing
over things. It’s chaotic but mostly works. These systems use “off-chain” governance in
which network participants “vote” by deciding whether and how to upgrade the software
they are running.
synctext commented 2 years ago

Discussed progress, DAO blog reading and Twitter checking.