shoo

67 ms·

1. 661.▲
   by shoo 5y ago

   Jennifer Senior's excellent article on burnout, "Can’t Get No Satisfaction" [1], contains this gem of a quote, from psychotherapist & burned-out burnout researcher Prof Barry Farber: > [...] their level of caring coul
2. 662.▲
   by shoo 5y ago

   I've avoided C++ for most of my career, but one idiotic mistake I do remember making was accidentally leaving an open curlybrace at the end of one of my source files. The C++ compiler ran and reported 1000s of compilation errors all t
3. 663.▲
   by shoo 5y ago

   Vector quantisation is discussed briefly in the Russell & Norvig section on speech recognition. I'll paraphrase: In speech recognition a sub-problem is to figure out how to encode the conditional distribution P(features|phone), whe
4. 664.▲
   by shoo 5y ago

   if you're willing to sink time into a lecture series, you may enjoy Richard McElreath's introductory lectures on bayesian inference: https://www.youtube.com/watch?v=_NEMHM1wDfI
5. 665.▲
   by shoo 5y ago

   > Theoretical grounds for applying the same update rules to loopy graphs were later developed that explain loopy BP as an approximate variational inference method in which inference is cast as an optimization problem. Instead of directly
6. 666.▲
   by shoo 5y ago

   maybe it's also partly to deal with the ongoing costs of customer support.
7. 667.▲
   by shoo 5y ago

   what would true agency actually look like? would the team "owning" the system get to capture the bulk of revenue or value the system produces -- so they could decide if they wanted to stack on another feature that might improve r
8. 668.▲
   by shoo 5y ago

   A pathway to understanding TrueSkill could be first getting an understanding of factor graphs and Minka's expectation propagation algorithm [1][2]. I only have a superficial understanding of those things, but I think I've got an u
9. 669.▲
   by shoo 5y ago

   here's an arbitrary review of Pearl's recent book, the book of why: https://www.ams.org/journals/notices/201907/rnoti-p1093.pdf
10. 670.▲
    by shoo 5y ago

    i agree. `find somewhere -exec some_command {} +` can be dramatically faster. but it does not guarantee a single invocation of `some_command`, it may make multiple invocations if you pass very large numbers of matching files after spending
11. 671.▲
    by shoo 5y ago

    What would be the motivation for people to get involved? Can they compete for riches? Can they become famous (with respect to the niche of prediction-brainstorming) by topping some leaderboard? Do they receive fractional IP rights and roya
12. 672.▲
    by shoo 5y ago

    Related -- there's a 2011 post from Paul Minero with fast approximations for logarithm, exponential, power, inverse root. http://www.machinedlearnings.com/2011/06/fast-approximate-lo... Minero's faster a
13. 673.▲
    by shoo 5y ago

    for R there is always https://ggplot2.tidyverse.org/ i am unsure if ggplot2 is in vogue, having not used R for a decade
14. 674.▲
    by shoo 5y ago

    (this is also not my field) I was thinking about this a little more: In some cases (such as your excellent wug example) it might be possible to heuristically associate the word "wug" to a probable contextual meaning "your cat
15. 675.▲
    by shoo 5y ago

    > A machine does not do that; instead it'll give words intrinsic, non-negotiable meanings. It won't have "wug" in its dataset, so it won't understand that sentence. > I don't believe that machine learning
16. 676.▲
    by shoo 5y ago

    tangent: one of the more satisfying patches of my programming career was improving system stability and roughly preserving performance while removing the use of parallel processing the code grew like this: 1. some python tool is written to
17. 677.▲
    by shoo 5y ago

    There's a variant of this in $enterprise-dayjob: production system Foo is a service exposing a FooAPI with a spec documented in some machine readable format (SOAP ; Swagger/OpenAPI ; gRPC ). You don't need to infer the spec
18. 678.▲
    by shoo 5y ago

    accepting PRs would have the downside of generating additional work for maintainers of the repo to review PRs.
19. 679.▲
    by shoo 5y ago

    > Rama has ten mangoes with which he wants to earn fifteen annas. Krishna wants only four mangoes. How much will Krishna have to pay? If Rama has a transactional perspective on business and has the local monopoly on mangoes, and Krishna
20. 680.▲
    by shoo 5y ago

    i agree that the task of actually searching for a valid proof or counterexample of some hypthosesis (rather than verifying a given purported proof is correct or not with respect to some known axioms/theorems etc) relies on an awfully l
21. 681.▲
    by shoo 5y ago

    i am not a beatles fan, i have no idea. presumably they played it live (or the component parts of it) during recording. if they never played it live, it makes the idea of regarding the live sound as the "true" sound even more absu
22. 682.▲
    by shoo 5y ago

    > the sound that was recorded and post processed at the time and released is the correct one that's a fun angle - what is the "correct" sound? a different way to define the "correct" sound could be to claim that
23. 683.▲
    by shoo 5y ago

    Seems like a hackathon. s/hack/work/ . Compared to the $1m netflix prize from 2006, $3.5k seems like a meagre bounty. There may be the case where someone has already been thinking hard about solving similar problems for the l
24. 684.▲
    by shoo 5y ago

    (I've been trying to understand variational bayesian inference): tangent: One application of information geometry pops up when trying to do gradient ascent(descent) to maximise(minimise) some function over a space of probability distri
25. 685.▲
    by shoo 5y ago

    > We have shown a proof in ACL2(r) of the Fundamental Theorem of Algebra.[...] The proof follows the 1746 proof of d’Alembert, which was maligned by Gauss on the grounds that it was not sufficiently rigorous. Gauss was right, in that it
26. 686.▲
    by shoo 5y ago

    nikola's demonstration truck prototype was "powered" by gravitational potential energy, not by hydrogen.
27. 687.▲
    by shoo 5y ago

    tangent: For Hindenburg Research's business model, it doesn't suffice for them to accurately detect that a company is fraud. Hindenburg also needs to get the timing of their trades right, and find a way to communicate the message
28. 688.▲
    by shoo 5y ago

    I completely agree. A thriving business with a poor quality codebase is the problem you want to have. Use the cash generated by business to pay developers to carefully refactor the offending parts of the codebase or architecture in a direct
29. 689.▲
    by shoo 5y ago

    I reckon that all things equal, bad code quality in itself does not increase the chances of success. Bad code quality may be a symptom of rapid iterations and not burning too much cash on development prior to achieving product-market fit. A
30. 690.▲
    by shoo 5y ago

    For a different (no doubt biased) perspective, see how Chinese companies listed on US stock exchanges are perceived by many US short sellers. One sentiment I have heard from US short sellers is that if Chinese companies defraud US investors