j7ake

24 ms·

361.
by j7ake 3y ago
People are no longer looking for ML , they’re looking for LLM.
362.
by j7ake 3y ago
Because you’re assuming just from the name or photo thay someone knows the language and have job permit. This means children from immigrant families growing up in Europe would be discriminated against simply because of how they look and wha
363.
by j7ake 3y ago
If you could choose any time period (and assuming median standard conditions, so cannot choose being king), which period would you think was best for raising children?
364.
by j7ake 3y ago
Any data to support this hypothesis?
365.
by j7ake 3y ago
Companies should rather ask the candidate to describe a project they completed. Let the interviewer then dive into any technical challenges such a project may propose, in order to vet technical skills. This allows a chance for candidate to
366.
by j7ake 3y ago
Your answer does not address the time component. Take an example: an electronic system is designed to give a solution to the traveling salesman problem on a list of a billion of destinations. You turn the system “on” so it can start doing t
367.
by j7ake 3y ago
But how long of a time frame are you allowed before you make this judgement on what a system does? Clearly any sufficiently new system will do “nothing”, because not enough time has passed to produce results.
368.
by j7ake 3y ago
I think what it is is whether the person feels accepted in their “group”. It becomes more painfully obvious that they aren’t accepted when surrounded by a group of people versus being alone.
369.
by j7ake 3y ago
What you state do not seem like priors at all. They look like observed data. A prior would tell you how much each of those attributes would change your viewpoint. Having very strong priors actually means your decision won’t change unless yo
370.
by j7ake 3y ago
They’re usually first year graduate students. Or fresh out of undergrad.
371.
by j7ake 3y ago
Get into a PhD program that is quantitative or mathematical. You’ll pick up coding on the way.
372.
by j7ake 3y ago
I predict the cost to make an 80 percent solution will plummet, which is great for consumers. Essentially, consumers now might pay X amount to get only a 40% solution (probably unusable and not something a company would even offer), but wit
373.
by j7ake 3y ago
If music is mathematical, then what isn't mathematical?
374.
by j7ake 3y ago
It is also true that a clunky and inelegant piece of code that does its job is infinitely better than a pristinely written code that solves a slightly wrong problem. So iterating quickly and being agile is bad for the code quality, but it i
375.
by j7ake 3y ago
“We choose to [have kids] and [have a career] not because they are easy, but because they are hard. Because that goal will serve to organise and measure the best of our energies and skills,”
376.
by j7ake 3y ago
It’s also not meaningful if that career coerces you to give up your outside family life (by imposing a huge penalty for having kids) so you can keep working your desk job, report to a boss, and have your 3 weeks of vacation a year.
377.
by j7ake 3y ago
Housing in Tier 1 cities still extremely unaffordable though.
378.
by j7ake 3y ago
How long does it take to really “read” GEB? There are puzzles and problems inside that could really extend how long you spend in that book.
379.
by j7ake 3y ago
Do you also attend music concrete? Or attend a conference? They are analogous. You go to watch people you admire do what they’re good at, in a crowded atmosphere.
380.
by j7ake 3y ago
Because academics care more about which journal a paper is published in, rather than the content of the paper.
381.
by j7ake 3y ago
Cash only restaurants are often 10-20 percent cheaper than comparable restaurants.
382.
by j7ake 3y ago
It's interesting how they say that their model has high "zero-shot" performance, even though it was trained on millions of examples. To me, zero-shot would mean a model whose parameters was never tuned through training exampl
383.
by j7ake 3y ago
I found three things very useful about ChatGPT: (1) A Google Search replacement. (2) An excellent text summarizer. (3) A quick prototyper (code and math).
384.
by j7ake 3y ago
How would binomial work here? You’re choosing amongst 101 urns of different proportions. How do you narrow 101 choices into two?
385.
by j7ake 3y ago
As long as you think the flipper chose their coin from a bag of coins with varied biases, then first flip always gives actionable information. It’s mostly irrelevant what kind of distribution this bag of coins comes from (except for some de
386.
by j7ake 3y ago
The problem being discussed is the same as my toy example, just extended to more coins. You now have 101 coins, with biases ranging from 100 percent heads (or urn with only red balls) to 100 percent tails (or urn with only green). Someone c
387.
by j7ake 3y ago
How would normal distribution change the result? I am imaging a random number generated from truncated normal distribution with mean 50, clipped 0 and 100: the answer will still lean right. How would binomial work? You have 101 urns from wh
388.
by j7ake 3y ago
You have two coins, one is biased heads and the other biased tails. Someone picks a coin without you knowing which one, then starts flipping. The first toss shows heads. Does that mean subsequent flips from the same coin will be more likely
389.
by j7ake 3y ago
The solution is to abandon quantitative university rankings. They're obviously not useful.
390.
by j7ake 3y ago
You're almost there with your intuition. You just now need to simulate drawing two balls without replacement for each of your 101 urns (from first urn with n=0 to last urn with n=100). Then you take the conditional probability of seein

More ›