It baffles me that people actually take these assertions seriously, especially after having used different software that uses voice input, like Siri, Google Assistant, Alexa or whatever. Those things make some serious mistakes even under ideal circumstances, and you want me to believe that they can accurately overhear things in non-ideal circumstances? I highly doubt it.

Regardless, you can use an ad blocker to make this a moot point - I’ve never experienced anything even close to this, because I never get ads.

I’m sure they make enough money to not care. Being in the part of the company that brings in the dough is generally a pretty good position to be in as well.

Should have say: self-hosting is always superior to cloud hosting.

That statement still comes with a pretty damn big caveat though - you need to have the know-how, the time to invest and the hardware (i.e money) to actually set something like this up.

If all of those are true, then self-hosting can definitely be an attractive option for you.

It’s only true for a vanishingly small fraction of the population, though.

Hence, Bitwarden is a pragmatic solution that will be superior for the vast majority of the population.

Also, local software and database is always superior to cloud.

Now there’s an unfounded blanket statement if I ever saw one.

Bitwarden is probably a more pragmatic choice for most users, given that it’s free and without having to manage the syncing yourself.

Any password manager is better than the alternative, though.

Product launches are the vehicle for attaining promotions at Google, allegedly. Maintenance does not get similarly rewarded, nor does launching projects and having them live on to actually be successful.

When the launcher got promoted and moved on, they have to figure out whether to keep the thing around, and the answer is generally going to be no since few things can really compete with the infinite money glitch that is search ads.

The cost of storage in this case is more or less irrelevant - traffic is what matters here. You’re also not getting any mentionable bulk discount on the servers for that matter.

The key is that you can engineer things in completely different way when you have trivial amounts of traffic hitting your systems - you can do things that will not scale in any way, shape or form.

Their scale was also an insignificant fraction of what Netflix has, making the point even more irrelevant.

The best figure I could find on Jetflicks user count was 37k, where as Netflix has 269 million users.

I don’t see any good reason why the merits of hydrogen for vehicle fuel would be any better than production and disposal of batteries. The other cases I agree that hydrogen will have a useful niche.

It’s hard to assess the validity of those claims as the article doesn’t bring any numbers and the paper itself is paywalled. As the fossil fuel industry is pushing hard towards wedging in hydrogen as a means of keeping themselves alive for a while longer, it’s vital to be able to assess the actual claims, lest they are just planted there by the fossil fuel industry.

There are some use-cases where hydrogen will be useful, but I don’t think storage is one of them. Nor do I think vehicles are a particularly good use-case either, as compared to just iterating on battery technology.

and is a good way to store excess energy from solar and wind.

Is it really that good of a storage method, though? The round-trip efficiency is quite bad when compared to other methods of storage.

I mean, not really. This is actually a non-trivial topic, and true random is a really bad label for what someone actually wants out of a shuffling algorithm.

See the following engineering blog post on the subject: https://engineering.atspotify.com/2014/02/how-to-shuffle-songs/

Data takeouts are non-optional under the GDPR, so I would be very surprised if that happens.

True random shuffle would be a terrible idea. No one wants the same track showing up multiple times in a row, which would not be uncommon in true random shuffle.

I’ll keep it, cheers

The article is talking about killing it from a product usefulness-perspective, not a monkey making-perspective.

More powerful hardware makes tasks that were previously not considered end-user tasks feasible for end-users, just give it some time.

Growing up, there was an association in my area for common ownership of different types of machinery and other equipment for its members. You paid something like $10 a year, and for that you got to borrow all kinds of things you might need as a home owner, like a wood chopper/splitter, high pressure washer, trailers, leaf blowers, cement mixer, scaffolding etc.

I always thought that was brilliant.

https://www.levels.fyi/t/software-engineer/levels/senior/locations/san-francisco-bay-area

vs

https://www.levels.fyi/t/software-engineer/levels/senior/locations/munich-metro-region

Even if you assume that additional labour costs are a bit higher in Germany, there’s no way on earth that it could explain the difference of the Bay Area-median being over 3 times higher.