l-lousy

It says “Uber and Lyft”, I highly doubt he’d be allowed to code at both.

I feel like the title should add “ in bats”. With all the new variants being discovered in people (Delta, Omicron, etc) I initially thought this was going to be the source of my next 2 week stint in lockdown. Seems like this is more of a warning against a new version of MERS which kills so fast that it doesn’t spread far.

Well he’s (Sam) very likely still living at home, so parents are paying for most things. Plus he has the revenue from YouTube

So they’re doing both masking for the inputs and knowledge distillation? Is it the combination of these two methods that’s novel?

I think an average user password would be shorter than a sha256 hash no?

I think they offer very similar things. Games offer you the ability to interact, making the story at hand more your own. Books give you less information, allowing you to create the environment for yourself, but have a set story.

There are also games to teach you things (factorio definitely taught me a lot about planning, not to mention logic gates), just as books can.

The only advantage books have for me is that they don’t irradiate my eyes with as much blue light as a screen would late at night.

He mentions using Rust/Rocket and SQLite in the article, I think it was just focusing on the front end library

Did nobody read he was using rust/rocket and SQlite for the backend?

You have one of the most wonderful looking blogs I’ve ever seen!

Why isn’t it possible for both groups to have an equal work life balance?

Holy crap that’s amazing! How complex is your model? And are there lots of parallelizable parts like filters or is it recurrent?

He started the article by saying that there were already enough articles written about the positives. While it does sound a little like he’s not happy with his job, I think he acknowledges off the bat that there are a lot of positives as well.

Not if, as R2 says, they only charge for objects that exceed a single digit request per second threshold. Your write once read once model still only costs as much as the storage at rest would.

Yeah I agree with rust, I have some simple database fetching servers that run crypto functions as well (I.e. login) and for simpler operations they can quickly get to 100K+ requests per second on a single host

I find that it’s generally the sender that has trouble serving 1GBPs, I’ve occasionally gotten it from my AWS instances when the result is in memory already

No I have the same. It’s regular priced plan, discounted by half for everyone in the building for the first two years. There are no ways in the contract to raise the price before that.

Don’t use JS

I’m saving up for the flood of old GPUs. I can finally realize my dream of taking all my ML work off the cloud once I scoop up some old mining GPUs for relatively cheap

I would use one as a source of universal online ID in account recovery, not for everyday use

Just keep all our two-headed alien friends off the ring world and we’ll be fine

Look up the ring World Series by Larry Niven

This felt somewhat wrong to me (as I’m sure it did many people that have take a Physics class) and ended up finding a Reddit discussion about this[1]. Seems like it’s not totally correct because the function for the ball becomes discontinuous

[1] https://www.reddit.com/r/Physics/comments/mn11r/the_dome_a_s...

I think there are efforts to create oracles like LINK, but you’re right that the only trustworthy data is from within the chain.

Based on another link posted here it looks like the hackers just called the half of a cross chain transaction that does the Eth payout. Could they not just make sure that there is a corresponding TX in the other half of their handshake rather than trusting the caller is actually their other contract

Very unlikely to succeed if you choose someone with enough vested interest in the company. Also “Dave is the only one that could have approved this, he’s who we sue”

They can do the same thing the hackers did and launder it before withdrawing to a new wallet

This is mainly discussing WFH, so in this case people actually are living in the office.

Interesting article, this also seems like a form of knowledge distillation. There have been a lot of examples of people distilling an ensemble into a single model, maybe you could try that here directly by taking out the middle man (match their outputs directly instead of labeling data).

Anyway, I’ve been trying to think of how this could be used for text data, specifically NER, which generally requires a lot more semantic understanding of the input. Sadly it seems like there might not be much room for the ‘micro’ part of the micro models.

The 4th dimension is used as more of a metaphor or minor plot point. The series has a much more valuable message.

The counterpoint here is that the internet archive isn’t a for profit service which is claiming exact reproductions of other people’s content as “new work”