36 Matching Annotations
  1. Jan 2021
    1. Because the cost of compute resources on the Internet Computer can be made approximately constant

      providing ICPs ultimately depends on markets: if you need to pay in advance for a piece of software that is compute-heavy, wouldn't you need to potentially market-buy a whole lotta ICP to be ready? --> price increase --> non-constant price for compute

    2. by enabling hosted open internet services to publish non-revokable, “permanent” APIs. Essentially, this involves their developers marking shared functions as permanent, which ensures that if their controlling governance systems try to push a software upgrade that would remove or change them, it will automatically fail.

      hmmm...

    3. The point is that if you build on the APIs of Big Tech’s services today, you are certainly building on sand, but even when you build using the APIs of those who aren’t monopolies in the making, you still run an enormous risk that one day they will decide to demand payment, revoke your access owing to strategic considerations, or simply fail. Consequently, it has become increasingly difficult to raise money to develop internet services that depend on shared data and functionality, and while the tech world becomes ever more monopolistic, innovation and economic opportunity increasingly suffer, and the original dream of a programmable web is dying on the vine.

      facts.

    4. Secondly, although users must identify themselves to the smart contract systems they are interacting with using a cryptographic key pair, management of such keys is made far easier when compared to the use of traditional blockchain wallets. Either user experiences can have users identify themselves by entering traditional usernames and passwords within the browser, which are deterministically converted into a key pair for them by code in the browser, or far more securely, and even more conveniently, they may be allowed to authenticate themselves using the emerging WebAuthn standard. This allows users to login quickly using the secure hardware features of modern client devices, such as by simply pressing the fingerprint sensor on their MacBook laptop, or authenticating themselves to their phone.

      this is interesting. what are the security implications? why don't we have webauthn for ETH contracts yet?

    5. it might be that an electro magnetic pulse caused by a solar flare from the sun disables some of the hyperscale data centers in which they live, or something else entirely; there are many different ways it could go catastrophically wrong

      if that ever happens, we got bigger problems

    6. a regulator provides an excuse to switch them off

      facts... though not a complete parallel, governments can shut off power to independent datacenters running nodes as well...

    7. Consequently, we can surely say that the practice of using Big Tech cloud services to build public blockchains and related systems and services, which are meant to be open, unstoppable, and tamperproof, is folly and antithetical to this purpose. The Internet Computer provides a solution.

      Agree with the former, but not necessarily with the latter...

    8. Thus, the Internet Computer network prevents a handful of clouds becoming intermediaries for compute power. Instead, it runs exclusively upon dedicated node hardware installed in high-quality, independent data centers around the world, and can withstand corruption and failure across multiple geographies and jurisdictions.

      This isn't really a valid knock on PoS networks; after all, they too can spin up data centers/hardware and run node software, effectively hedging against potential downtime of big cloud providers.

      Who would you bet on? AWS or relatively smaller datacenters for 99.99% uptime?

    9. Most importantly, of course, it is the tamperproof nature of blockchain that makes it possible to support autonomous open systems, and DeFi, using smart contracts. By extending the provision of security on blockchains to privacy, the design of systems and services based on smart contracts has been simplified and the scope of their application greatly extended.

      For better or for worse.. smart contracts come down to code.. and code is inevitably exploitable.

    10. aim to rebuild the world’s infrastructure and critical online services in unstoppable form on the Internet Computer as soon as practicable.

      Aren't nodes vulnerable to outages as well? E.g. who's spinning up the node infrastructure? If it's hosted on AWS as well, aren't we in the same boat?

    11. On the Internet Computer, however, you might solve this need simply by making a function call to an open internet service that provides arbitration and dispute resolution, which will later return the result with another simple function call.

      So it sounds like the argument for IC is that while this exact functionality is possible on existing protocols (ETH), it simply isn't scalable enough

    12. This means, for example, that if service A, which is built on the Internet Computer, shares functions with service B, the code of B can directly call the shared functions of A exactly as though it were calling its own functions, even through services A and B are actually instances of running software, rather than static libraries, and even when they are written in different programming languages — which is revolutionary.

      First of all, what is the need for this? Aka what are the limitations of using static libraries?

      Secondly, from a dynamic standpoint, how does this differ from smart contracts interacting with each other?

    13. One of the greatest challenges met is the provision of a unified, on-chain environment, that can process any volume of smart contract computations and maintain any quantity of smart contract data.

      word???

    14. Internet Computer — the world’s first public blockchain capable of expanding its capacity without bound, that runs at web speed

      primary explanation of their usecase

    15. dissolve their neurons to release the tokens inside, then convert them into cycles to power computation; or transfer them, as best suits their purpose

      explicit tx types?

    1. While his original major was in English, Harris became interested in philosophical questions while at Stanford University after an experience with the empathogen–entactogen MDMA (colloquially known as ecstasy).[12][13][14] The experience led him to be interested in the idea that he might be able to achieve spiritual insights without the use of drugs.[15] Leaving Stanford in his second year, a quarter after his psychedelic experience, he visited India and Nepal, where he studied meditation with teachers of Buddhist and Hindu religions,[15][16] including Dilgo Khyentse.[17] Eleven years later, in 1997, he returned to Stanford, completing a B.A. degree in philosophy in 2000.[18][19][20] Harris began writing his first book, The End of Faith, immediately after the September 11 attacks.

      ngl, straight up a fascinating journey

  2. Dec 2020
    1. It goes like this: to take the first step and test (validate) an idea, you don't need absolutely anything. None of these excuses-from-the-future: venture capital, developers and cool CTO, technology, "can you please recommend me an iOS designer?", money for marketing and buying traffic. Here's how it works:

      do unscalable things to validate an idea

  3. Nov 2020