Best Twitter thread of the day – August 17th

The Ethereum merge (switch to proof of stake from proof of work) is a few weeks out. For most folks in the Ethereum community, this is maybe the most meaningful event of the past five years.

Some advice for beginners (and experts):

— Robert Leshner (@rleshner) August 17, 2022

First things first, don’t fall for scams.

In times of great uncertainty, people question things less. Be extra frosty about what you read, click, and sign.

— Robert Leshner (@rleshner) August 17, 2022

Don’t use too much leverage, take too much risk, or do something stupid with your portfolio. This is not the time to get caught up in mania.

Volatility may be really high, in both directions.

— Robert Leshner (@rleshner) August 17, 2022

Make sure your assets are custodied safely.

That could be on an exchange you trust, or on a hardware wallet you control.

If you haven’t thought about security lately, now is a great time to revisit it.

— Robert Leshner (@rleshner) August 17, 2022

If Ethereum forks (such as a proof of work chain splitting off), the “state” of Ethereum won’t transfer with it.

Every stablecoin and DeFi application on the fork is likely to fail immediately.

You wont suddenly be double wealthy.

— Robert Leshner (@rleshner) August 17, 2022

Be extremely careful about signing transactions on a fork chain; you have one private key that works on both chains, and a miner could potentially replay a transaction on Ethereum.

Don’t “sell all my worthless (POW) USDC for a little bit of (POW) Ether”.

— Robert Leshner (@rleshner) August 17, 2022

If you’re not sure what to do, do nothing. Seriously.

— Robert Leshner (@rleshner) August 17, 2022

.@balajis “just make papers function calls” has a bit of an “import antigravity” vibe to it. but i love the idea lol

random thoughts:

0/13

— L (@llllvvuu) August 17, 2022

1/13

if i’m writing a meta-analysis and i import a few studies as .csv. i have no reason to expect that the column headers will match, the units will match, etc. i will probably have to impute some data and come up with a covariance matrix / hierarchical model, etc

— L (@llllvvuu) August 17, 2022

2/13

but this is reasonable, ofc each paper’ll have some business logic in addition to importing other papers, and then export the cleaned .csv along with pickled priors, model specification, pickled model weights, maybe some recursive snarks (for fast verification), etc…

— L (@llllvvuu) August 17, 2022

3/13

but it gets murkier when it comes to “vision papers”, papers which are mainly interpretation/philosophy/normative/etc. maybe these aren’t supposed to he covered by the framework. or are we supposed to reject these from the scientific canon? make scientists learn coq or die?

— L (@llllvvuu) August 17, 2022

4/13

there’s not much we can do to verify the integrity of data besides sign where it came from and have a diversity of sources and include fraud risk into the priors of meta-analyses. but how many research questions warrant this much budget?

— L (@llllvvuu) August 17, 2022

5/13

probably need to improve the pipeline efficiency if we want these high-quality outputs

definitely “citation as function call” hides lots of complexity and will actually require much infrastructure, as well as retraining (pain) for scientists, which might dip throughput/$

— L (@llllvvuu) August 17, 2022

6/13

if we want to have a dependabot for scientists then we should distinguish between critical dependencies and “mentions”, kinda like how npm has dependencies and devDependencies. but if we’re ambitious we could even try granularly tracking each export to a set of imports

— L (@llllvvuu) August 17, 2022

7/13

i think also that the citation graph doesn’t have an obvious reason to live on-chain yet. i think the files can just be

“`
import <contenthash> as a
b = businesslogic()
export b as c
“`

which is perfectly legible off-chain

— L (@llllvvuu) August 17, 2022

8/13

we can use $FIL, $AR, or even USD to incentivize mirroring the global repository

blockchain shows up only when we need to resolve double spends (e.g. revocation). so can be very minimal bytecode / storage on-chain if any at all

— L (@llllvvuu) August 17, 2022

9/13

compute-to-data / FHE stuff is quite exciting for consensually collecting encrypted data from civilians. it would also be quite bonkers for the ML people to figure out data cleaning, since it makes up most of data analysis. but that’d also be “the singularity” probably

— L (@llllvvuu) August 17, 2022

10/13

two things which i don’t think are safe design:
* automating reward splits based on citations
* automating rewards based on test suites passing or w/e

in both cases, recommendations/notifs are useful but human dev/scientist hybrids should determine and sign off on rewards

— L (@llllvvuu) August 17, 2022

11/13

tl;dr i think “citation as function calls” roadmap is something like:

* being (jupyter/r markdown/etc) notebook maxis (kaggle kekw) with a standard import/export format (i.e. language interop) based on contenthash (nix kekw)

— L (@llllvvuu) August 17, 2022

12/13

* recursive snarks so the import tree is verified w/o traversing
* revocation chain for retractions
* retraction propagation / fraud proofs / dependabot

— L (@llllvvuu) August 17, 2022

13/13

it sounds mostly like “digitize science” but one could argue that “blockchain” and “decentralization” are also just part of the digitization journey

— L (@llllvvuu) August 17, 2022

‘Proposal for Expedited Referenda’ has been posted to the Acala community governance forum. The proposal suggests changes to time periods in Acala's governance process and outlines next steps for Acala.

Details in thread⤵️https://t.co/Lwh2ISJyq0

— Acala (@AcalaNetwork) August 16, 2022

2/10

All decisions made for Acala's network are done via community-driven, on-chain governance. On-chain voting occurs before any changes to the network can be executed. On-chain governance is a feature of Acala's parachain thanks to @Polkadot & @substrate_io.

— Acala (@AcalaNetwork) August 16, 2022

3/10

As an Acala voter, please consider doing the following:

• Engage in voting so we can collectively move forward

• Help sense check the proposals

• Help verify the details of the proposal

— Acala (@AcalaNetwork) August 16, 2022

4/10

The following will come as next steps over the coming days & weeks:

A series of trace reports will be published detailing transaction traces for each of these 16 identified addresses. The community can collectively formulate proposals to resolve the error minting of aUSD.

— Acala (@AcalaNetwork) August 16, 2022

5/10 – What's next (cont.)

Further trace reports will be published to identify arbitragers profiting from aUSD error mints, as well as outflows of aUSD error mints and tokens swapped from aUSD error mints to other accounts, parachains, and exchanges.

— Acala (@AcalaNetwork) August 16, 2022

6/10 – What's next (cont.)

The community can then collectively formulate proposals to resolve the error minting of aUSD and outflow of funds.

— Acala (@AcalaNetwork) August 16, 2022

7/10 – What's next (cont.)

Analysis of the liquidity pools and relevant activities will be published. The community can collectively formulate proposals to handle the states of the pools.

— Acala (@AcalaNetwork) August 16, 2022

8/10 – What's next (cont.)

A post-mortem report will be formulated and released with a summary and improvements to be made going forward

— Acala (@AcalaNetwork) August 16, 2022

9/10 – What's next (cont.)

Analysis of the state of the paused protocols and operations will be published. The community can collectively formulate proposals to gradually resume normal operations.

— Acala (@AcalaNetwork) August 16, 2022