tromp

897 ms·

2971.
by tromp 6y ago
In some settings, "repeated addition" is the only sensible way to think of multiplication. Such as multiplying an elliptic curve point with a scalar.
2972.
by tromp 6y ago
Wrong; there is a large backlog of transactions paying over 50 sat per byte [1], which is over $10 in fees for most transactions. [1] https://jochen-hoenicke.de/queue/#BTC,24h
2973.
by tromp 6y ago
The point remains that eliminating the backlog of high fee paying transactions would jeopardize mining stability and security in the long term.
2974.
by tromp 6y ago
> simply transition to proof As opposed to Proof-of-Work, Proof-of-Stake is not as simple as it sounds. Any implementation faces a myriad of design challenges and potential attacks. Things like "nothing-at-stake", "costles
2975.
by tromp 6y ago
The block subsidy is constant between halvings. And is thus a step function. But "emission" refers to cumulative block subsidies. And thus is piecewise linear, with 34 pieces of exponentially decreasing slope, the last one ending
2976.
by tromp 6y ago
> The total energy usage would be reduced 600-fold! It wouldn't. If X$ worth of bitcoin is handed out every block, then miners competing for that prize are willing to collectively spend up to X$ per block on average. If they could o
2977.
by tromp 6y ago
It's inverse exponentially approaching the 21 million limit, which it can never reach [1]. [1] https://medium.com/amberdata/why-the-bitcoin-supply-will-nev...
2978.
by tromp 6y ago
With an emission of 1 per second forever, Grin is more like time (I make the case for uncapped supply in [1]). [1] https://john-tromp.medium.com/a-case-for-using-soft-total-su...
2979.
by tromp 6y ago
Satoshi was using "node" in the sense of mining node, as is clear from "The rest will be client nodes that only do transactions and don't generate." So he imagined mining to move to large datacenters, and users runn
2980.
by tromp 6y ago
a majority of hashpower can enforce a softfork (tightening of rules), but has nothing to do with a hardfork (change of rules). If non-miners stick with old rules, then miners using new rules simply get ignored by everybody else.
2981.
by tromp 6y ago
You need to verify the history as well and continually update the UTXO set. That takes a lot more time, especially with a spinning hard disk, since the UTXO updates are not particularly sequential.
2982.
by tromp 6y ago
You start with {x, y, z=0, a, b=0} and then reversibly compute z+=x; z+=y, b+=z; b+=a to end up with {x, y, z=x+y, a, b=x+y+a} from which you could reversibly compute the starting state again with b-=a, etc.
2983.
by tromp 6y ago
Running a fully verifying node doesn't require that much space, since you can prune away older blocks once verified (if you're fine with not being able to seed other nodes). It does take a fair amount of time to download so many G
2984.
by tromp 6y ago
> The original plan for Bitcoin envisioned larger blocks to keep transactions fees down. The plan was always for security to be provided by transaction fees in the long term, as block subsidy dwindles to insignificance. But you just can&
2985.
by tromp 6y ago
Same here on MacOS Catalina 10.15.7. The original invocation runs fine $ bash -c './redbean.com -vv' W2021-02-26T01:50:12.665277:tool/net/redbean.c:1144:redbean:3302] setsockopt(server, IPPROTO_TCP, TCP_FA
2986.
by tromp 6y ago
I expect transaction fees will dominate the block reward within 4 halvings, as they are already within an order of magnitude of the block subsidy...
2987.
by tromp 6y ago
What led framework to choose Intel CPUs over AMD ones? Did it simplify the design of the system?
2988.
by tromp 6y ago
Relatively straightforward with bitcoin, but not entirely trivial, since there can never be more than 20,999,971.02187096 bitcoins. The simplest possible emission is 1 per second forever.
2989.
by tromp 6y ago
except that with difficulty, you know the expected effort needed to pass the threshold. with folding of arbitrary proteins, you have no idea.
2990.
by tromp 6y ago
You can verify that some fold achieves some energy state. What you cannot efficiently verify is that it's the best possible fold, achieving the global minimum energy.
2991.
by tromp 6y ago
And my point was that once fees dominate the block reward, the price doesn't need to double with every subsidy halving for mining income to grow. A smaller increase in price will suffice.
2992.
by tromp 6y ago
No, fees will be $20-$100, similar to now. Block subsidy will be 16x lower than currently, so block reward will be mostly fees.
2993.
by tromp 6y ago
ChessBase just posted a new article by A. Silver at https://en.chessbase.com/post/how-a-neural-network-is-made where the author tries to argue that doubling the neural network size in StockFish makes for an appreciable
2994.
by tromp 6y ago
> as long as the price of bitcoin outpaces the halvings Not needed, since within 4 halvings, we can expect transaction fees to dominate the block subsidy.
2995.
by tromp 6y ago
You can rephrase your question as: how many $$ have been spent on mining bitcoin (both capex and opex)? Although answering that question still requires assumptions about miners selling bitcoin to cover their costs...
2996.
by tromp 6y ago
Spot the error in this sentence: "By the end of 2018, according to some estimates, miners here could account for anywhere from 15 to 30 percent of all bitcoin mining in the world, and impressive shares of other cryptocurrencies, such a
2997.
by tromp 6y ago
Airdrop by captchas is miles less egalitarian than making people mine for coins in a fair competition over decades. Worse yet, there's no way to verify that a majority of coins didn't end up in the hands of just a few people
2998.
by tromp 6y ago
The cost of diskspace is not as limiting as the time needed to do an Initial Block Download, which would become too burdensome with huge blocks. Transacting on 2nd layer makes a lot of sense as txs are instant, more private, and don't
2999.
by tromp 6y ago
> Bitcoin is reproducing the same type of wealth distribution we see in real life. That would be the best it can hope for. But it's very far from that. Just consider that one person (Satoshi) holds ~10% of all bitcoin. Or 2 other pe
3000.
by tromp 6y ago
It is 70% premined. And its enormous complexity combined with frequent consensus model changes makes for a much larger attack surface than the simpler, close to immutable, bitcoin.

More ›