woko

As a Colab user, I have got to admit that I am impressed. I have used your website today to give it a try, and the interface is responsive (and thus somewhat addictive).

I use GPU on Colab, so I will stick to Colab for now, but I think I will hop on Deepnote from time to time.

I would not say that there is no point trying to memorize a full book. It is an inefficient use of one's time, sure. However, I believe it can have some benefits. Typically, the process of trying to memorize (assuming one does not do it in the dumbest way) leads to a deeper understanding of some subtlety of the subject.

I think this post means the same as this tweet:

> Transformers are a special case of Graph Neural Networks. This may be obvious to some.

https://twitter.com/OriolVinyalsML/status/123378359362695168...

As others mentioned, it is a good habit to request the page to be archived. You don't have to link to the archive, but you would have the option to if the page were to disappear in the future.

I wish I had done this 15 years ago for a small project/website. Nowadays, my website is there, with all of its content, but most of the awesome references which I had linked to are unavailable. I wrote "most", but it is close to all of them.

> do anarchists really work as professors at the London School of Economics? And Yale?

I know a banker who is communist. It does not mean that he behaves like a communist in his actions, but he does in his words. From my perspective, he takes the best (or the worst) of both worlds: the quality of life of a very rich man who can afford living in the most expensive places and whatever leisure he wants, and the ability to punch down on the "lesser" people around him from his morally higher stand-point. Never under-estimate people's hypocrisy and cognitive dissonance, especially when what is at stake is exerting power over other people's lives. One can always find a way to rationalize one's crappy behaviour to try to lower the cognitive dissonance after the fact.

> we’ve never moved permanently backwards before, so it would be rather strange if we started doing that now.

It would not be strange. It would be unexpected. It reminds me of this funny little story:

> "Every morning on a turkey farm, the farmer comes to feed the turkeys. A scientist turkey, having observed this pattern to hold without change for almost a year, makes the following discovery: “Every morning at eleven, food arrives.” On the morning of Thanksgiving, the scientist announces this law to the other turkeys. But that morning at eleven, food doesn’t arrive; instead, the farmer comes and kills the entire flock."

I have the same issue. I never waste time trying to find a way to forwards these emails to the right people, because I am extremely suspicious, so I often assume there is a phishing attempt behind these emails. I mean, often, money is mentioned in the email (like booking a hotel, a bill, a car insurance, Internet subscription, etc.). It might be legit, but it is not worth the hassle for me, and it is none of my business anyway.

The issue is that new competitors have to go against... Apple and Google. Given what happened to Epic (retaliation on Unreal Engine after #FreeFortnite), you can be sure that any competitor which would get a little bit of traction would be hammered into the ground by some Apple/Google money-powered shenanigans.

> What data and models were used to synthesize these predictions?

You have to ask the creator of ArtBreeder (https://www.artbreeder.com/about) which apparently uses BigGAN, StyleGAN, etc., trained on different datasets to offer these options:

https://i.imgur.com/5jTNgXl.jpg

That being said, there is a heavy manual touch (based on "Photoshop and historical references"). See for instance how different Septimus can appear compared to the webpage:

https://twitter.com/dvoshart/status/1290004339646517250

> I'd be careful with computer crimes on the Internet though.

Exactly. Let us say you break into a shop owned by some mafia to put out a fire, then you might be fine w.r.t. authority, but you might be in trouble w.r.t. criminals. Similarly, say you break a car window to pull out a pit-bull left alone in the sun, you might have some issues with the owner if he turns out to be part of some drug trafficking gang.

There is no reason to believe that phishing websites are run by script-kiddies, there are obviously criminal rings running all sorts of businesses on the Internet too. I would rather leave the work to the authorities rather than risk going through trouble with unknown criminals, just so that I could have my funny revenge over them.

The webpage is weird. For these interested, it seems to be about:

https://github.com/DanielBMarkham/OnePageChess

The videos are very impressive. I wish you let us move the 3D scene with the mouse in the browser to get a better idea of the result.

Not a doctor, but I have had kidney stones.

The answer is a loss of calcium, so high level found in urine test: if you had way too much vitamin D, then you would start releasing the calcium from your bones.

I don't think that taking calcium supplements would help, because calcium metabolism (https://en.wikipedia.org/wiki/Calcium_metabolism) is much more complex than that.

I would call it "inheritance" of discrimination rather than "replication" of discrimination: the skin-colour bias arises because of the Ivy-League bias.

Quotes below:

> "[...] Facebook staffers say that despite adding more schools to the recruiting lists, White managers continue to select from the same Ivy League and West Coast schools they’d attended.

> Many of these institutions act as another racial filter because Black students are often under-represented.

> “Companies will give unfair weight to a school like Stanford, dipping down to the top 20% of students, but might pass on a student in the top 2% of their class at Penn State [...]"

As a non-native speaker, this website is useful to learn English slang.

I get your point, but I don't get how it diverges from what was described in the article. Based on what you wrote, it looks to me that you are actually very advanced in the process of forgiveness: telling your story, expressing your feelings, seeing your childhood through the eyes of yourself as an adult, focusing on your well-being in the present and in the future.

> they keep digging up the past and weaponizing your own failures against you in everyday fucking conversation.

I know that. You have no idea how much your post hits home. This kind of behaviour is exactly what I try not to replicate, because it is the best way to dig one's own grave. There is nothing good which can come out of it.

I think you have an issue with the word "forgiveness", and maybe it is not the right word. What I understand from its use in the article is being able to let go the hatred (way stronger than anger), and being able to focus on yourself in order to improve your well-being and to bring positivity into your present and future life.

True. I realize that I wrongly assumed that "revenge" was retaliation perpetrated by the victim outside of any legal system.

If I understand right, some people consider that justice is a form of revenge. If the process of administering justice is done by a legitimate third-party, which lets the victim focus on healing/repairing, then I agree that justice will help the victim.

You kind of answered the question: revenge does not help the victim. It might feel necessary, but it does not actually help.

Listen to some samples:

- https://openai.com/blog/jukebox/ (2020, quite good, but no classical music)

- https://openai.com/blog/musenet/ (2019 so not as good as the 2020 one, but showcases classical music)

There is no reason to assume that one cannot be moved by AI-generated music, as the AI has learnt from human-generated music and tries to mimick the styles.

> I'm not aware of parents that specifically believe learning music makes their kid more intelligent in non-musical areas.

It is all anecdotal evidence, but I am aware of some parents who "forced" their baby (and then little kid) to listen to classical music (specifically Mozart), because the kid would become a genius. Apparently, after some googling, it is called the "Mozart effect". Here are press articles about it:

[1] https://www.scientificamerican.com/article/fact-or-fiction-b...

[2] https://www.bbc.com/future/article/20130107-can-mozart-boost...

Could it be a motivation for teaching music to little kids? That is what is suggested in the conclusions of the articles linked above, and which I quote below. Some parents might be receptive to these conclusions.

From the Scientific American article [1]:

> Rather than passively listening to music, Rauscher advocates putting an instrument into the hands of a youngster to raise intelligence. She cites a 1997 University of California, Los Angeles, study that found, among 25,000 students, those who had spent time involved in a musical pursuit tested higher on SATs and reading proficiency exams than those with no instruction in music.

From the BBC article [2]:

> There is a way in which music can make a difference to your IQ, though. Unfortunately it requires a bit more effort than putting on a CD. Learning to play a musical instrument can have a beneficial effect on your brain. Jessica Grahn, a cognitive scientist at Western University in London, Ontario says that a year of piano lessons, combined with regular practice can increase IQ by as much as three points.

Disclaimer: I only realize now that my comment is about images returned in image search results, not about reverse image search.

Bing definitely uses facial recognition: faces of celebrities are detected, there is a bounding box around the face, and their name and a link to Wikipedia. It does so for clothes as well.

Here is an example (fully automatic):

- the two faces: https://i.imgur.com/bfVNvW6.png

Here is another example, where clothes and items are automatically found in the image, and a bounding box is automatically defined:

- the necklace: https://i.imgur.com/vfhvY0T.jpg - the suit: https://i.imgur.com/AqS3qBV.jpg - the other suit: https://i.imgur.com/jTgRwpb.png

However, for the faces in this second example, I have to draw the box myself:

- the face: https://i.imgur.com/Pe1RLLm.jpg - the other face: https://i.imgur.com/LHOOwnW.jpg

So, what do you mean? What makes Yandex special with respect to facial recognition?

For me, it is a setting to tick, which mentions the country:

https://i.imgur.com/fML3x0l.png (in French)

Once it is set to my country (France), then I get almost identical results.

I don't know how to deactivate this feature on Bing, so I could not compare Bing and DDG in the case where the "location" setting would be ticked off on both sites.

In my opinion, it is due to location. Bing returns 8 results on the first page, and DDG returns 10 results. At the top of the Bing page, I can see that Bing already takes my location into account (and it is precise, it is my city), so there is no box to tick. Once I tick the "location" box on DDG (it is only the country), I get 7 identical results out of 8 on the first page! The order is the same for results 1-4, the 5th result is different, the 6th result is identical, and results 7-8 are identical but their positions are swapped.

It is troubling that there was nobody in the meeting room to raise concerns about pushing notifications to ~250 millions[1] of users.

[1] Monthly Active users at https://data.firefox.com/dashboard/user-activity

Thank you. I appreciate your conciseness.

The presentation is similar to the (nice yet more verbose) notebook "01-g-h-filter.ipynb" in github.com/rlabbe/Kalman-and-Bayesian-Filters-in-Python/

I find it weird that this (free) book was omitted:

- Goodfellow et al. Deep learning. MIT press, 2016. https://www.deeplearningbook.org/

And there is the book about DL with Python published by Manning in 2017, but not the book about DL with PyTorch, published by Manning in 2020?

- Stevens et al. Deep Learning with PyTorch. Manning Publications, 2020. https://www.manning.com/books/deep-learning-with-pytorch

From your link:

> The way this process works is that, while averaging all of the pixels in a series of, say, 10 images, the program mathematically calculates which pixels fall far away from the mean value because they're much brighter (or much fainter) compared to the same pixels in other frames. The algorithm then discards those out-of-range pixel values so they don’t affect the final image.

Wouldn't this process remove part of the comet trails as well as the satellite trails?

I mean, I get how it works if all you care about is relatively static like distant stars, but would it work for this specific use case?

> generating a barrage of spam or fake news that uncritical readers can't distinguish from human output.

Which could indirectly bring good things: what if this decreases our tolerance threshold for nonsensical texts?

For instance, people might become more impatient when faced with click-bait articles, because they know they might have been written by an AI in order to waste their time. This would let people skip the articles after the first sentence without feeling any guilt or doubt about it.

At least, I hope it might open the eyes of uncritical readers.

> I briefly skimmed (not expecting very much from a Bitcoin-themed article), read the end, and only then read more carefully.

It sounds to me like you must be an academic, or someone with good habits for being efficient at reading articles.

Same, except that I skipped straight to the last line to check whether it was a generated text after I noticed the first sentence made no sense (GPT-2 already generates grammatically correct English sentences).