Lecture 2 - How Bitcoin Achieves Decentralization

Second lecture of the Bitcoin and cryptocurrency technologies online course.
For the accompanying textbook, including the free draft version, see: bitcoinbook.cs.princeton.edu/
In this lecture (click the time to jump to the section):
* Centralization vs. decentralization 1:12
* Distributed consensus 4:45
* Consensus without identity: the block chain 17:46
* Incentives and proof of work 35:43
* Putting it all together 55:39

[@DaveDavis1]

That was absolutely fantastic. Aravind is a really great lecturer and clearly practiced this well in advance for our enjoyment. The subject matter is perfectly explained. This series so far is pure awesome. I'm so glad I waited until they were all released, I can't stop watching!

[@silent7152]

Are you rich now

[@samir_485]

I'm so thankful we get to watch this for free. This series is just amazing and well made.

[@Joshua-tg6hj]

Brilliant lecture. Very informative; the best one I have found.

[@MrGribouillage]

Really clear, I heard and read a lot on bitcoin, nothing was as clear as this !

[@SV-tc8cu]

Wow , this is. goldmine. I tried MIT and Stanford lectures on crypto and they are garbage comparing to this. Thanks a million.

[@Chopik1149]

Excellent. GREAT series. So well presented. Much appreciated. Thank you.

[@JoukoSalonen]

great lecture again. Learned a lot about consensus (without identities). Thank you.

[@MrRigdensChannel]

Best video of the series

[@piggubiggu5324]

Can't believe my eyes. Such a brilliant lecture 👍

[@johnmacdonald4560]

Thank you for the video, amazing job by the lecturer!!!

[@pazuzil]

Thank you for such an amazing lecture!

[@orbea1951]

that's a great lecture at any level, thank you

[@likkiii07]

Beautiful Lecture!

[@gusgonzal3z]

Great lecture!

[@TheBerkobe]

First lecture was a little bit more "technical". This one is easier to understand but I had to watch some parts again and again. Hope this lecture series will help me to understand my investments on cryptocurrencies. Thanks for your precious effort!..

[@kirchdubl1652]

What is a mechanism of deleting transactions from the poll if they made to the valid block ?

[@kirchdubl1652]

can some nodes broadcast , that there are 2 transactions which are specious (after all there is a big chance that all nodes see 2 transactions from the same address while block is mined, do quick check and reject it - or broadcast that these 2 trans are malicious ?)

[@hushi3403]

this is awesome for bitcoin novice

[@ramsaybolton966]

So what Silicon Valley TV series picturised about 51% attack is all wrong. Rules of network cannot be changed??

[@matrixstuff3512]

Why was such a simple explanation of blockchain/bitcoin so hard to find?

[@TheBitcoinArmy]

nice!

[@manonlescaut4701]

22:53: Does every node lists (almost) every outstanding tranactions or is there sort of a distribution of the outstanding transactions among the nodes?

[@ianmathwiz7]

Manon Lescaut Nodes keep and broadcast every transaction they see that isn't already in the blockchain, and doesn't conflict with any transactions they've already seen. The exception is that if they see a transaction that is identical to an existing pending transaction, but with a higher transaction fee, they will use the version with the higher fee.

[@noahrodriguez1895]

Lot of information. but towards the end you start understanding it as it all gets put together.

[@ccerrato147]

I'm currently studying cryptography and distributed systems algorithms in detail. What else should I study to get to eventually master the concepts and algorithms used for cryptocurrencies and bitcoin inspired systems?

[@ccerrato147]

+georgem No, sorry. I'm just in Coursera for Cryptography and Cloud Computing which shoes P2P systems. Maybe the authors of the video could help you?

[@danielr4115]

Awesome lecture. But I have a question: If you wait for n Blocks to be passed, that contain your transaction, and the probability of double spending decreases exponentially, there is a probability of double spending (independent of n). From a purely mathematical point of view, considering the huge number of transactions made, that would imply that there were a finite number of double spending attempts that actually succeeded, and this number is growing. Doesn't than the probabilistic argument imply that the Bitcoin- system is inherently faulty?

[@heruka111]

One of the chains is going to be dropped and one of the transactions will be considered failed. If Bob waits at least 5 blocks he is nearly guaranteed that one of the chains has dropped and can either consider the transaction a success or failure, but no double spending has occurred.

[@youbaji]

if bob has to wait 6 blocks to confirm and each new block takes 10mins to create, then the transaction speed is 1hr. Alice has to wait 1hr to download a software. Too slow.

[@SV-tc8cu]

Looking at other sources, they mention 2140 as a year when last bitcoin would be mined, not 2040. Is there a typo at 40:14 ?

[@youbaji]

why higher computing power node not aways win? if searching algorithm brute force searches space in given order, faster computer will always discover search space faster and therefore always win.

[@kirchdubl1652]

it is like saying, if you buy $100 worth lottery tickets makes you always a winner assuming everybody else is buying only for $1.

[@forheuristiclifeksh7836]

18:47 nodes are malicious

[@felixkookie]

Should it be more precise? I love precision.

[@ataggake]

Great work, but there is a rather huge error at 41:00. Block reward finishes around 2140, not 2040.

[@JoukoSalonen]

he may have a point with that, not sure though, check this Adam: bitcointalk.org/index.php?topic=235926.20

[@ataggake]

Jouko Salonen If you're referring to the graph, no it's misleading. The rewards form a geometric sequence (of the type 1 + 1/2 + 1/4 + ... = 2). The sequence would continue to infinity except for the fact that once the reward drops below 1 satoshi it's rounded down to zero. The fact that the size of the reward is therefore decreasing exponentially from one 4 year period to the next causes the graph to look like it ends a few decades from now (because the size of the reward has become orders of magnitudes smaller). The real end of block rewards is around 2140 and nowhere near 2040.

[@1st-Impressions]

Jouko Salonen Technically the math says 2140, but practically bitcoins will be mined out by 2040.

[@ataggake]

Derek Watson Practically, they will be mined out by 2038 or 2032 or 2065. It just depends how you zoom in on the graph. The only meaningful *specific* time is the time when the reward drops from 1 satoshi to 0 satoshis, which is around 2140.

[@1st-Impressions]

Adam Gibson I think 2040 is a generally accepted figure, in the same way that 6 confirmations is quoted in the video as providing the same security that only an infinite number of confirmations would do in your universe. :P

[@jmcdonne]

Minor correction at 52:56 : the first word of Poisson process should be pronounced pwa-sawn , not poy-zone.

[@DANS-tj3df]

Wait..... Total # of bitcoin runs out in 2040???????? 🤔See min 40........

[@shymaaarafat1342]

-I kind of "not convinced", or have questions, about some issues: 1-Why do u accept a Nonce that leads to just "less than"(

[@turnkit]

This video appears to be possibly deinterlaced with the wrong field order. For whatever reason there is some weird jutter going on.

[@DaveDavis1]

It's because they (thankfully) fade in and out between the slide and the side on the screen in real life. This way, we can see what he's pointing at with his laser pointer. They forgot to do it on one slide in the first lecture and it was super annoying so I'm glad they didn't miss any here. This is by design, not a technical fault.

[@qitianlong1110]

why does it half every year?

[@lawrencemwangi8846]

every 210000 blocks ~ 4 years

[@ianmathwiz7]

It halves every four years. The reason is so the currency doesn't indefinitely inflate.

[@SoCalFreelance]

You can't decentralize BTC software development which is a flaw. We have to trust core developers. Only a few people in the world have the technical competencies/time/motivation to verify/validate their work.

[@2minvideos175]

hi, are you from india?(tamilnadu)

[@chrystophercaserodizon3189]

COLLISION=BIGBANG=(/C.E.R.NCDC5#).

[@silvertassel9383]

Learned a lot. Thank you! ***** 35 cents *****

[@victorcontreras9725]

I'm from the future, 2019. Ummm.... bitcoin is now just store of value... you will experience a crash in 2018.... next bull run will start in 2019. ...that is all...

[@leomoraes9]

This guy should have some lessons with AAntonopoulos.

[@samnada6248]

51% Attacks would break crypto 😎

[@unutilizzatoreyoutubbicoca7749]

😎

[@hardefakta4469]

Bitcoin is not decentralized. It's highly centralized

[@trifalgarh]

Explain how?

[@JoaoSilva-uk9gl]

Because it's controlled by the big miners... And this a very bad sign because at any point in time they can simply reverse the game and go from honest to attackers.

[@trifalgarh]

Ow okay, thanks :)

[@turidtuva3012]

Except its not controlled by them.. The whole segwit ordeal proved this! And if they were to attack, they could do nothing to the already recorded history on the blockchain, they could only reject new transactions, and open up for double spending. However, if they were to do this, we would for sure find out, and the value of bitcoin would absolutely plummet, and a 51% attack would therefore be a bad decision, as they now sit on all this expensive hardware, and the cost of power of running such rigs would be higher than the reward.

[@yerawizardaiden]

One sentence : Go through the lecture again. Also here, we might want to look at what centralisation and decentralisation mean again, so that we don't mistake what we perceive to be a "concentration of power" to be exactly the same as centralisation. Any chance that it may end up being centralised is nullified and also they do not control which transactions go through or that they cannot ensure that the transactions that they deem to be invalid never make it onto the ledger and they can never control the system like Scrooge in terms of ScroogeCoin. Also Turid Tuva's argument further sums it up perfectly. Plus I don't think I need to go into any further unnecessary details other than what has been covered in the lectures so far.

[@alexl.349]

After all, I don't like the lecture. Pace is too high even for me and I have PhD in math.

[@Daniel-hi8wc]

Pause button + Google

[@SV-tc8cu]

There is no correlation between having PhD in math and your ability to follow the explanation of this lecture. This is not a math lecture.

[@daitavan297]

Boring lecture !