63 Matching Annotations
  1. Last 7 days
    1. Why doesn’t Anki use SuperMemo’s latest algorithm? The simple answer is that SuperMemo’s latest algorithm is proprietary, and requires licensing. As Anki is an open source application, it can only make use of algorithms that have been made freely available. We’re inclined to believe SuperMemo when they say their newer algorithms are more efficient, but feel that to a certain extent, it is a case of diminishing returns. The gains achieved by moving from a traditional study routine to SM-2 are already great, and by sticking with an open algorithm, your learning data is not locked into a single product. Ultimately it’s up to you to decide - if access to the latest and greatest scheduler is a higher priority than the things that Anki brings to the table, you may want to check out SuperMemo to see if it is a good fit for you.

      Most SRS programs use SM-2 because the latest is proprietary (and SM-5 had issues in Anki).

      Any benefit you get from using the latest SuperMemo is small compared to the benefits you might get from sticking to an open platform (exponential growth). Most of the SRS benefit is already gained from moving to SM-2.

  2. Jun 2022
    1. probably one of the most important things about Zettelkasten is that you’re always in search for related or already existing notes. because reusing components is more efficient reusing components is more efficient so we don't need to "reinvent the wheel" related to [[reuse code for efficiency and maintainability]]... 6/20/2022 expanding a new note can be dangerous because it might take the atomicity out of the note making it harder to reuse the component. reusing components is more efficient reusing components is more efficient so we don't need to "reinvent the wheel" related to [[reuse code for efficiency and maintainability]]... 6/20/2022 . So it’s better to keep notes smaller and contained

      It is better to link notes than to add to notes. This improves reuse and is easier.

  3. May 2022
    1. Taleb is suggesting that miners might simply stop supporting the network, especially when the block reward ends (around 2140). Without miners supporting it, bitcoin’s value could drop to zero. His assumption is that if bitcoin has no revenue and a theoretical future value of zero, it should be considered as having no value now. He ignores the fees that miners receive for processing transactions alongside the Coinbase reward, which will rise exponentially if the network effect of usage continues. That is a big if, and the behaviour of miners after the 21 million hard cap is unknown, but that doesn't necessarily equate to zero value.

      Tail emissions (like Monero has) may fix this.

      An exponential fee rise would probably be bad for the network, especially when there are alternatives.

    1. Along with this, perhaps one of the biggest advantages of exploring data through a career at a firm like Citadel is that you receive instant, objective feedback from the markets and can continuously evaluate your work and professional development.

      Instant feedback like an interpreter - learning can be addictive.

    1. But on average this isn’t possible. Moreover, it’s very hard to predict which startups will succeed, and easy to be overconfident in your skills. In hindsight being employee 100 at Facebook was a great idea, but how would you know at the time? Only a handful of VC funds seem to be able to pick winners with any amount of regularity, and even if this is due to skill instead of luck, the fact that the vast majority of experts whose full-time job is to choose winning startups repeatedly fail at this task should make us wary of assuming an engineer with minimal business understanding could succeed.

      It's fairly easy to beat average. The bell curve is very skewed towards passive job hunting.

    2. The average developer in Mountain View makes $106,000 per year,4 so the early startup employee has a 24% edge. However, it’s likely the early employees work harder, and are be more skilled than average. Rather than comparing averages to averages, it seems more accurate to compare startup employees with a more experienced subset of developers. A level II developer at Google in the Bay Area makes $150,000 per year,5 while a senior developer at Google makes $250,000 per year.6 From this perspective, if you’re turning down a job at a top company in order to work at a typical startup, then you can expect to earn less than similarly qualified and hardworking peers.7

      This doesn't even count the equity the big company engineer earns.

    3. For a “reasonable” person, it is better to be an early stage employee than to be a founder. This fits with advice given by people like Dustin Moskovitz.

      If you have enough money then it could be great to be a founder.

      If you aren't reasonable/rational, maybe it's worth simulating one (i.e. following this advice blindly).

    4. On average, startup early employees earn at most only a little more than developers at larger companies.

      Does this apply to CEO/CTO as well?

      The risk/reward tradeoff might not be worth it.

      Might be another labor of love like a PhD (which is beneficial, but not for the cost)

    1. Financial surveillance: Oleg's parents send him some Bitcoin to pay for textbooks, then continue to snoop on his Bitcoin address and activity. A few months later, Oleg sends some leftover Bitcoin to the public donation address for an organization that does not align with his parents' political views. He does not realize that they are still monitoring his Bitcoin activity until he receives a furious email from his parents, berating him.

      Is this what dusting attacks are?

    1. Building on DB, Auth, and our compute infrastructure we want to continue investing in simple yet scalable components that let people build and compose software seamlessly. Our north star vision is to shorten the distance between idea and product to its absolute essence. By thinking of a good idea, you should be half the way towards realizing it.

      Most people probably dont want to trawl the internet to find out the best database to use, etc...

    1. Back to keeping everyone in a multiplayer repl in sync. In the real world, we don't always work at the same time. Sometimes people leave then rejoin later, or lose their internet connection and come back online. When people rejoin, we need their client to "catch up" to the current state of the repl so everyone has the same code in front of them!

      Can you play back what they were doing? It'd help me so much to see what my team has been changing. Especially seeing a time-scrub when doing a PR review...

  4. www.fooledbyrandomness.com www.fooledbyrandomness.com
    1. There is a mistaken conflation between success for a"digital currency", which requires some stability andusability, and speculative price appreciation.

      Isn't this constant(ish), governed by the difficulty?

      I'm not sure I trust LN but thats an alternative. Or XMR, or 0 conf - it seems as secure as a bank transaction (I think?) - probably good enough for coffee. For big purchases, wait the ten minutes.

    2. We exclude collectibles from that category, as they havean aesthetic utility as if one were, in a way, renting themfor an expense that maps to a dividend — and thus are nodifferent from perishable consumer goods. The same appliesto the jewelry side of gold and its convenience yield: my goldnecklace may be worth 0 in thirty years, but then I would havebeen wearing it for six decades.

      why doesnt this apply to the short term utility of bitcoin? (of which there isnt a whole lot... lets say monero instead)

    3. Earnings-free assets with no residual value are problematic.The implication is that, owing to the absence of any explicityield benefitting the holder of bitcoin, if we expect that atany point in the future the value will be zero when miners areextinct, the technology becomes obsolete, or future generationsget into other such "assets" and bitcoin loses its appeal forthem, then the value must be zero now3

      Why must it be zero now? If miners are expected to be extinct in 20 years, but it serves a purpose now (assume it does), then there is some value between now and then.

      FAANG stock is expected to be zero in 500 years - should they be zero now?

    4. A central result (even principle) in the rational expectationsand securities pricing literature is that, thanks to the law ofiterated expectations, if we expect now that we will expect theprice to vary at some point in the future, then by backwardinduction such a variation must be incorporated in the pricenow.

      this sounds a lot like the efficient market hypothesis, which I do not believe (see Dr Daniel Kim lectures)

    5. Cryptocurrencies require a sustained amount of inter-est in them

      Precious metals, fiat and stocks do too... (real estate as well but those serve a tangible purpose)

    6. Gold and other precious metals are largely maintenancefree, do not degrade over an historical horizon, anddo not require maintenance to refresh their physicalproperties over time.

      You do not need software to store gold. It's harder to shoot yourself in the foot. Gold wallets are cold wallets by default.

    1. Your first sentence needs to grab the reader. Go back and read my first sentence to this post. I rewrote it a dozen times. It makes you curious. That’s the key.

      Make sure you get the first sentence down. The rest is important but not as important.

    2. Business writing is about clarity and persuasion. The main technique is keeping things simple. Simple writing is persuasive. A good argument in five sentences will sway more people than a brilliant argument in a hundred sentences. Don’t fight it.

      Short is persuasive!

    1. Actually, he launched “Neoreason” to build a suite of products that augment human reason, starting with programming, but quickly adopted the name Replit from the open source project when they realized how hard it would be to build just that one product really well.

      Wow. This is what I've been wanting to do...

    2. I view entrepreneurship as means of reconciling the infinite and finite. You venture into your imagination, gather knowledge, and dream about a better world. But you have to bring some of that back to earth. You take a step -- no matter how small -- towards your imagined world in the real world.

      There's that twitter quote about how the quickest way to stop being a perfectionist is to start a company...

    3. The beauty of React Native is that it lets developers write code in React, a best-in-class JavaScript library, that automatically works across platforms. That means that small companies might not need an iOS team and an Android team and teams for the long-tail of platforms. One mobile team shares the same codebase, and React renders the code to the native platform’s UI. One entrepreneur I spoke to this week brought up React Native unprompted, saying, “With tools like AWS, React Native, and more, we can do an enormous amount with a team of 10 people.” 

      Pattern: Lowering barrier to entry. Leveraging talent of fewer people.

    4. In talking to Ryan, I realized we had the same vision of what we want to do, of how we want to make the programming accessible to everybody. We had similar thoughts about coding and what we call “⁠the new literacy⁠”⁠ and how everybody should learn how to program. 

      Do I really believe everyone should be able to program?

      They should have the logic ability at least... and computers are such powerful and potentially low-barrier tools that it could benefit the human race A LOT if the skill were normalized.

    5. A truly open, decentralized internet will be one that a billion people can build on and remix and compose, where anyone can interact directly with protocols, APIs, and services, and build protocols, APIs, and services themselves

      Composability is essential here, as even with a platform to make coding simpler, it would be truly accelerated if you can just compose larger concepts together.

    6. But one thing that still hasn’t been solved is that most of us aren’t technical enough to actually create software. Fewer than 30 million people in the world know how to code. Anyone can now own the protocols and apps they use, but most of us (certainly me included) still can’t build them, or build on top of them.  

      Only 30M people can code?!

    7. If you can acquire the youngest users, retain them as they get older, and continue to attract the new cohorts of young users, you will win over time. 

      Kids are the biggest and most expansive demographic to acquire users from.

    8. Where everyone can manipulate code like we manipulate word

      Reminds me a gain of Roam.

      What if humans spoke in compilable code? What if we thought in that? What if we do?

    9. Remixing = breeding of memes (in the Richard Dawkins sense).

    10. shared data

      NFT access to data?

    11. snap together composable services

      This is cool. Would also be awesome if you could contribute your own compute.

    12. disruptive innovation

      Why haven't other repl platforms been disruptive so far? (or have they?)

    13. Actually, I’m just going to call it: Replit will be a $100 billion+ company, but more importantly, its users will generate multiples of that number in value for themselves. 

      What are other companies worth? How realistic is this?

    14. Codecademy

      What did Amjad Masad do at Codecademy?

    15. Because of the emergent nature of composability, it’s impossible to predict what Replit will look like in a decade, but it’s not impossible to predict that the results are going to be special. If the company gets everything just right, I think that it has the chance to be a $100 billion+ company that reshapes the way that software is developed and by whom, and by extension, who can build the internet. 

      What if they aren't able to deliver on all of their vision? Like many many companies.

      If they solely become an accessible platform for people to learn programming (and easily take their projects off-platform), and improve the DX, that'd still be big.

    16. It also means that Replit needs to make it really easy to get started by putting everything you need to code something simple in one place, accessible from anywhere in the world with an internet connection whether you’re on a brand new MacBook Pro, a Chromebook, or even an Android phone. 

      It would make it easier for people to become teachers too (another way to make money)

    17. But when you’re setting out to build something as big as Replit wants to become, you need to start with a clear and ambitious proximate objective, something to tackle right now. 

      Replit has lofty goals, but at least they have a roadmap for them. The nearterm goals are also quite staggering though. Their growth at least seems good.

    18. That’s why Replit wants to be the first place that anyone writes a line of code. Then it can become the first place they build an app, the first place they collaborate on software, the first place they make money online, and ultimately, the place that they turn whenever they want to create software. 

      Helping people make money = BIG BUSINESS

    19. It wants to lead the transition from a stacked model of software creation to a networked one, and to create the world’s first massively distributed and collaborative operating system.

      Stacked model -> Networked model of software creation = think github as an IDE, anybody can code together, anybody can run anybody's code.

      The use of "operating system" is an interesting one too. It doesn't even have to be just for coding. Think Roam.

    20. It’s laying the foundation on top of which one billion people will build the future.

      In a gold rush sell pickaxes.

  5. Apr 2022
    1. TL;DR: Incremental Reading is about spacing your reading materials so you don’t necessarily have to finish reading in one setting.

      It reduces the cost of context switching, and in fact lets you reap the benefits of context switching. It encourages it.

    1. Privacy is not secrecy. A private matter is something one doesn't want the whole world to know, but a secret matter is something one doesn't want anybody to know.

      Privacy is the power to decide when and what is secret and who to.

    1. to develop concepts for thinking about how programmable attention systems can extend beyond productivity (which is where these systems are most prolific, as can be seen from above examples)

      Netflix-like autoplay but for binge reading...

      Although it is productive: Incremental Writing

    2. Programmable attention centers around making high-level decisions that orchestrate how we later spend our attention. It is rife in everyday life, albeit (usually) in a weak way. Examples include:email inboxes with snoozessetting a recurring event on the calendartodos which appear based on locationGetting Things Done “Tickler” Fileshabit-enforcing systems (e.g. Beeminder, StickK)spaced repetition (e.g. Anki, SuperMemo)ML-trained “smart email inboxes”

      We already use programmable attention in everyday life, but it should be more integrated, potentially at an OS level.

      HeyFocus and ad blockers could also be considered under this umbrella.

    1. The latest advances in machine learning — namely transformers and self-supervised learning in natural language processing — will also make it possible to build personalized discovery engines that organize and surface information that is timely, relevant, and impactful

      And possibly summarize it. And connect it to your own knowledge graph.

      Readwise's spaced repetition and integration with Roam/Obsidian is doing some cool stuff with surfacing connections and serendipitous reminders. Here is a Twitter thread I wrote to myself when I started playing with it: https://twitter.com/alexbowe/status/1476817961897783296

    2. collaborative Roam graphs

      <3

    3. serendipity that sends the user down a rabbit hole to find new ideas and the links between them

      Big Incremental Reading vibes

    4. between search and discovery (and sometimes browse) is a tension between two desires: speed to information versus exploration of information

      Recently I've been struggling with the fact that I want both physical and digital copies of books. Digital copies let me automatically draw my highlights together for remixing and sharing, but physical copies allow for serendipitous rediscovery. If I know I want to read a book but don't have time right now I often get it off the shelf and leave it somewhere where it will be a visual cue for me later. I have been wondering how I could simulate that electronically (I keep coming back to the idea of a spaced repetition/incremental reading like system at the OS level, to resurface windows/tabs/books/anything at a serendipitous moment. I also consider how those digital photo frames could maybe be used)...

    5. next wave of human-computer interaction will focus more on “pushing” us the right, relevant information versus the user “pulling” for information

      This is already happening with ads. It'd be nice if the same techniques were used to pre-emptively help us instead.

  6. Mar 2022
    1. If you’re worried about your reputation, publish pseudonymously. You’ll receive most of the benefits of writing online without risking your reputation. Plus, if you build an audience and change your mind, you can always come out from behind the curtain and start publishing under your real name.

      Like Sia and her crypto personality recently.

    2. Serendipity Vehicle” – a magnet for ideas and people and opportunities

      "Serendipity Magnet" (better term)

      Luck Magnet?

      Is cool how it attracts new ideas (and challenges and refines your own)

    3. Until recently, the average person wasn’t able to publish and distribute their ideas at a reasonable cost. But on the Internet, anybody, in any corner of the world, in any time zone, can access your best thinking. 24 hours a day. 7 days a week. 365 days a year.

      What were the cultural transformations that happened w/ papyrus / printing press / typewriter / internet?

      (from a lens similar to Third Industrial Revolution - would be cool to get someone to research this for me)

    4. Invest in your future. Whether you want to become an author or an entrepreneur, you should start writing online. No matter what kind of book you want to write or what kind of company you want to start, it pays to have a smart and targeted audience. If you plan to write a book, writing online will help you develop a voice and land a publishing deal. And if you plan to start a company, writing will help you build credibility, meet investors, and find future co-founders. People under-estimate how much work is done before you start a company: that’s when you build your network, learn what you need to know, and discover what you’re good at. The seeds are planted years before the flowers bloom.

      Set up feedback loops and build your network.

    5. When you create content, people can access your knowledge without taking your time. You no longer need to sell knowledge by the hour. Your ideas are the most valuable currency in a knowledge-driven economy. Just as an investment account allows your money to grow day and night without your involvement, content does the same with your ideas.

      People can also accidentally find it and share it - basically, people can do PR for you.

    6. Day and night, your content searches the world for people and opportunities. Projects, mentors, speaking gigs, job offers, pitches, investment opportunities, interview requests, podcast appearances, and invitations to special events. It all starts with sharing ideas online.

      It's like making your money work for you... passive investing.

    1. For me, my newsletter blurbs spark essay ideas and turn into longer pieces. For example, writing about Ron Chernow’s Alexander Hamilton turned into an inquiry on polymaths that I featured on the 39th and 40th editions of my newsletter. Eventually, these three editions turned into my essay on polymaths.

      Newsletter pieces can be recombined into longer pieces later.

      Flow: Twitter/conversations > newsletter > essay.

  7. Feb 2022
    1. Bisq has a very complex and resource hungry user interface. One of our main goals is to provide the user with a simple tool, that won’t need any technical knowledge.

      This really needs to be done on mobile.

      Not needing to run a node to be private, and mixing without a hardware wallet will be huge.

    2. The BSQ token is a huge privacy concern, when used to pay transaction fees on Bisq, it makes possible to link a transaction with a specific Bisq user.

      Dont use BSQ - it links BTC transactions to you.

    3. As a result of the points above: there have been reports of Bisq users seeing their accounts on centralized exchanges frozen after they sent BTC from Bisq to the account on the exchange. This means exchanges actively flag coins coming from Bisq as risky.

      This may not be true. Bisq being a DEX probably attracts coins that have been involved in darknet markets.

    4. Non custodial. The user has the total control of their wallet.

      Would be ideal to use any wallet (including hardware). We should stop implementing wallets specifically for every service.

  8. Jan 2022
    1. an Sperber, for example, uses the tools of epidemiology to study culture. In the broadest sense, it's the study of the patterns, causes, and effects of certain conditions within a population.

      Dawkins' "Memes"/memetic evolution

    2. How can 'mere' matter, properly configured, manage to be conscious? Are chimpanzees or elephants conscious? Can a computer be conscious?

      This reminds me of Godel, Escher, Bach.

    1. Highlights allow you to create markers in the video without getting distracted. When you click on the Highlighter icon, a note with the current timestamp of the video is saved privately to "Only Me". Highlights are of 10 seconds duration by default. Best way to use Highlights is to use them as placeholders for important moments in the video. Review them later, edit timestamps and save to appropriate Board

      Snippets and highlights seem basically the same other than the length.

      It appears that snippets are intended to be shared as their own video, whereas highlights are like chapters, or summaries for subsections. Although they look almost identical to me right now.

    2. A Summary Note is a special annotation that allows for quick timestamped note-taking on videos. Everytime you click on the Pencil icon, the note captures the current timestamp of the video. You can write your notes next to the timestamp. While viewing the video, you can click on these timestamps to quickly navigate to the point in the video that is refered by your note You can have only 1 Summary Note per video

      You can only have one summary note per video, so most likely this is used for summarizing the whole video.