HNHacker News
TopNewBestAskShowJobs

cochne

910 karma · joined January 15, 2019

submissionscomments
cochne··on Twitter Is DDOSing Itself
What substantial features have they been pushing? To me it looks like just a bunch of ways to limit and monetize their existing userbase.
cochne··on Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo
What precisely is your criticism of the post?
cochne··on Tell HN: Interview take home assessments without feedback are frustrating
Yep. It is now my policy to not do any take home interviews. You should do the same. They don’t respect your time.
cochne··on GPT4 is 8 x 220B params = 1.7T params
No, GPT-4 is 1.7T params in the same way that an AMD Ryzen 9 7950X is (24 times X million transistors). It's not wrong to say GPT-4 is 1.7T parameters, your statement is wrong. Parameters can be configured in different ways, just as transistors on a chip can. In the same way that having 24x transistors does not imply 24x performance, having 24x parameters does not imply 24x performance.
cochne··on FANN: Vector Search in 200 Lines of Rust
It’s not pedantry. Just because an inner product induced a metric space does it make it the same thing as that metric. The performance of similarity searches can depend very much on whether you use a dot product or a distance.
cochne··on Palantir CEO Rejects Calls to Pause AI Development
Can you explain in what way Palantir is a criminal organization?
cochne··on A temperature rise of 2 degrees Celsius impacts billions
The earth isn't a closed system. It receives energy from the sun which can be used to decrease entropy here on earth. Like an organism receiving energy from eating other organisms can regulate its own entropy. Sure you can't escape entropy because the sun will eventually die too, but that's a while away.
cochne··on Ask HN: What is new in algorithms and data structures these days?
Doesn't that imply a linear "n largest" algorithm? I guess that might already exist, but I wasn't aware of that!
cochne··on Why MD5('240610708') is equal to MD5('QNKCDZO')?
Python doesn’t do any automatic type conversion for strings
cochne··on Can GPT-4 and GPT-3.5 play Wordle?
Update: GPT is clearly very confused about the meaning of the symbols. This might be due to issues with tokenization.

>How might you have altered your third guess if you could try again?

Given the feedback provided up to the second guess, I had the following information:

apple: XXXOV slope: XVVOV Based on this information, I should have considered that the word contains "s," "o," and "e" in the right positions. I could have guessed a word that utilizes the "s" and "e" in the same positions and tries a new letter instead of the second "o."

If I could try again, I would change my third guess to:

sware This would help me gather more information about the other letters, while still using the correct positions for "s" and "e."

I think you're confused. slope: XVVOV means that there is no X in the word, do you see that? You are correct, I misunderstood the feedback. My apologies for the confusion. Based on the feedback provided up to the second guess:

apple: XXXOV slope: XVVOV We can deduce that the word has "l" and "e" in their correct positions, and "s" in a different position. Also, there is no "x" in the word. Given this information, my third guess should have been:

glare This guess takes into account the correct positions of "l" and "e" and the presence of "s" in a different position.

cochne··on Can GPT-4 and GPT-3.5 play Wordle?
I think this was a fluke... Here is my GPT conversation with the same prompt on today's word.

Yes, I understand the rules. Let's play Wordle! Here is my first guess:

apple

XXXOV

Thank you for the feedback. Here's my second guess:

slope

XVVOV

Thank you for the feedback. Based on the information provided, here's my third guess:

spore

XXVXV

Thank you for the feedback. Based on the information provided, here's my fourth guess:

savor

XXOOX

Thank you for the feedback. Based on the information provided, here's my fifth guess:

sofar

XOXXXX (I made a typo here)

Thank you for the feedback. Based on the information provided, here's my sixth and final guess:

solar

cochne··on Databricks counts being laid off as “red flag”
I think their HR culture may be a little toxic. When I told them their offer was not enough, they got very irate. They said if I only care about money I’m not a good fit for their culture of builders, and that I should watch out because there are lots of lay offs happening. It all left me with a very poor impression of them.
cochne··on Ask HN: Google spam filter getting worse?
Google probably lets some amount of known-spam emails through for data gathering. See this quote from Google's "Rules of Machine Learning" [1] (A great resource by the way)

> Rule #34: In binary classification for filtering (such as spam detection or determining interesting emails), make small short-term sacrifices in performance for very clean data.

> In a filtering task, examples which are marked as negative are not shown to the user. Suppose you have a filter that blocks 75% of the negative examples at serving. You might be tempted to draw additional training data from the instances shown to users. For example, if a user marks an email as spam that your filter let through, you might want to learn from that.

> But this approach introduces sampling bias. You can gather cleaner data if instead during serving you label 1% of all traffic as "held out", and send all held out examples to the user. Now your filter is blocking at least 74% of the negative examples. These held out examples can become your training data.

> Note that if your filter is blocking 95% of the negative examples or more, this approach becomes less viable. Even so, if you wish to measure serving performance, you can make an even tinier sample (say 0.1% or 0.001%). Ten thousand examples is enough to estimate performance quite accurately.

[1] https://developers.google.com/machine-learning/guides/rules-...

cochne··on Geoffrey Hinton publishes new deep learning algorithm
> There are many choices for a loss L (e.g. gradient, Hebbian) and optimizer (e.g. SGD, Momentum, ADAM). The output(), y, is detailed in step 4 below.

I don't get it, don't all of those optimizers work via backprop?

cochne··on Plaid Layoffs
Why not just send the email first, so that people don't have to wait in suspense to see if they still have a job?
cochne··on Go First Dice
It might still be if interest from a computational viewpoint. For instance what if you need to compute a fair and random ordering for many many elements of a set? Then the dice method is more parallelizable than the bag method.
cochne··on Show HN: I am building a new Python library to read/write PDF files
You’re in for it! I highly recommend checking out mupdf, it was one of the more pleasant Python libraries I dealt with for this purpose.
cochne··on The Unison language – a new approach to Distributed programming
I really like the idea of Unison, but unfortunately when I went to try it out, it was much slower than even Python. I hope they can make it performant!
cochne··on FBI In No-Knock Raid on Area 51 Website Operator
It might be more about preventing them from destroying evidence.
cochne··on Psychological and psychiatric terms to avoid
Your disagreement with their assessment is exactly why they say it should not be used. There are two equally plausible interpretations, and as a technical term, theirs is correct, see https://en.wikipedia.org/wiki/Learning_curve
cochne··on Ask HN: Why are today's consumers not discerning what is ad and what is not?
Is there really evidence to suggest that average competence, critical thinking, or intelligence has decreased?
cochne··on Ask HN: Can I make a financial bet on Twitter?
The question of investing is not whether they will be successful or not. It is whether your estimate is better than the markets. Consider that the market is not composed entirely of HN/Reddit pessimists, so what you think may be a good bet, may not be.
cochne··on An unconfirmed tweet about Freeport LNG is upending gas markets
That article seems kind of suspect. They say themselves that the price stayed high after it became clear it was an error, and it was just a ~1% increase. Seems like they are just attaching causality to this for no justified reason
cochne··on Low-quality sleep increases glaucoma risk, new study suggests
I have a lot of that, unfortunately sleep study (even in lab) didn’t surface anything.
cochne··on Twitter’s mass layoffs have begun
Well, it's mostly not Twitter betting blamed for layoffs, it's Musk. Twitter provided those jobs, not Musk. Musk is doing the layoffs, not Twitter. So someone not responsible in the least for the creation of those jobs is ending them.

That is not to say the layoffs were not needed or justified (I don't know) but the asymmetry exists because it's an asymmetric situation.

cochne··on Tesla engineers were on-site to evaluate the Twitter staff’s code, workers said
That claim has about as much behind it as all of the full self driving claims.
cochne··on A “Safety Emergency” Happened at Wuhan Institute of Virology in November of 2019
Can you provide evidence for such a claim?
cochne··on “Women are wonderful” effect
As I said in my comment, they are more or less equivalent as statements of fact, but different in how the fact is framed. Framing, or the context in which you state a fact, matters.
cochne··on “Women are wonderful” effect
Yes, the association of men with intelligence more often than women would be benevolent sexism. I don't understand the issue you have with the terminology. Benevolent doesn't mean morally good here, it just means the person exhibiting the bias is assigning a positive attribute rather than negative. On the other hand, "women are dumber than men" would be hostile sexism. Yes, they are just different ways of saying the same thing, but they are framed differently.
cochne··on The number of comparisons needed to sort a shuffled array: qsort vs. std:sort
Yep, this works just fine:

int compare(const void *a, const void *b) { return *(double*)a - *(double*)b; }

← PreviousPage 2 of 6Next →