simo7

Slavery. For instance the technology for windmills already existed at the time, it just wasn't that big of a deal in world with an abundance of slaves. Fast forward to the Middle Ages and you find it everywhere.

Interesting, I’m starting to think undocumented thresholds are quite common in GCP.

I experienced something similar with Clod Run: inexplicable scaling events based on CPU utilization and concurrent requests (the two metrics that regulate scaling according to their docs).

After a lot of back and forth with their (premium) support it turns out there are additional criteria, smthg related to request duration, but of course nobody was able to explain in details.

I have a different approach: I'm ok to throw away significant chunks of code as long as I have important new ideas that allow me to do more with less.

Often the problem with refactors is that don't come with enough of these new ideas , there's no real progress, it's mostly moving things around.

Fewer than 10k wouldn't be surprising at all, yes.

The 65k figure is what ancient historians reported, in reality it’s almost certainly order of magnitudes lower. Exaggerated figures are usually the case with ancient reports (especially about battles).

You can ship your local docker context to a remote host and build/run your containers there. All the docker commands you typically run locally you can run on the remote host.

https://www.docker.com/blog/how-to-deploy-on-remote-docker-h...

Isn’t a monolith with split workloads just a monorepo?

The main flaw of the article is not controlling for product category.

I suspect most data warehouses have similar NDRs.

In many companies a data warehouse is the place where you dump all your data and let everyone run poorly written programs against it.

Add to that poor engineering culture in data teams (often lead by non-technical people) and costs are bound to skyrocket.

I think the main lesson should be not to use it, especially at scale.

Not only they were cousins, in their private correspondence right before the outbreak of WWI they were addressing each other as Willy and Nicky[0]. By the way they fought against each other.

0: https://en.wikipedia.org/wiki/Willy%E2%80%93Nicky_correspond...

edit: I meant the correspondence between the third cousin, Wilhelm II, and Tsar Nicholas.

I think you're too quick at making assumptions which are not correct unfortunately. Let's recap:

there's no selective pressure for it to become less lethal.

If people take measures to prevent spread of a disease that’s literally selective pressure.

How trying to slow down its spread can "force" it to become less lethal?

The only possible answer is that lower lethality gives it more time to spread around right?

It turns out Covid has a long incubation period, in fact plenty to use the host to infect other people.

Tl;dr: Given the length of the incubation period a less lethal virus would not gain much in terms of its ability to spread. That's the thesis I'm hearing.

Agreed, macro predictions are very hard. Much easier to make predictions on individual assets (companies, commodities etc.) and hold long term.

Also, precisely because macro is hard, these analyses often feel superficial.

Inflation, especially if exogenous, can negatively impact the economy but at the same time cash-alternative assets become more attractive. What's the ultimate effect there?

And what about historically low interest rates? Don’t they warrant a shift in investment preferences towards stocks?

Wether a macro prediction turns out to be right or wrong it’s rare to read a deeper argument than “things are too high must go down”.

Don't really get your point.

The people have dramatically altered their behavior on a global scale (on average anyways) because of COVID’s lethality. There is absolutely selective pressure in that dimension.

This reads to me: "Covid is certainly more lethal than the average flu, therefore is undoubtedly getting less lethal" (effect of selective pressure). It doesn't make much sense.

there wouldn’t be anywhere near the amount of research into medical intervention

Right, how's that has to do with whether there's selective pressure or not?

It almost feels like you are attributing a different meaning to "selective pressure", but cannot understand which one.

Well if the two events (improved transimissibility and improved lethality) are not independent as I assumed then it's certainly bad news.

I also heard the opposite thesis btw, higher transimissibility might sacrifice something in terms of lethality. Like you can't really add much as I'm not an expert.

Yes my point is, when you do improve one (transimissibility) you probably did not improve the other one (or even got worse at it).

True, there's no selective pressure for it to become less lethal.

But I wonder: isn't a random mutation unlikely to improve on any of its profiles (transmissibility or lethality)? If so a more transmissible virus is unlikely to also be more lethal.

What if (not sure) a random mutation actually tends to do slightly worse? Then more transmissible viruses tend to be less lethal over time.

Sure we still get more lethal mutations from time to time, but the more transmissible ones are the ones which win out.

Well bitcoin does have an intrinsic monetary policy despite not having a monetary authority.

It is possible to engineer currencies to be inflationary/deflationary without an external authority.

It just cannot be a guarantee as there are many other factors to take into account.

even Bitcoin experiences inflation as more of it is mined (as does gold).

Surprisingly they don't understand what inflation means despite having just quoted its definition a few paragraphs above.

Inflation doesn't mean that we "have more of something", it means we "need more of that something" to buy the same good over time.

Do we today need more bitcoins to buy a pizza compared to 2013?

Nobody chooses to be a fool

Dogecoin is at least honest, so many other "jokes" including Bitcoin, contemporary art, some stocks etc...

If people want to play a game of musical chairs, I rather prefer this game to be called exactly what it is.

It surprises me to see so many warnings about Dogecoin and so few about Bitcoin for instance. People actually believe there's any relevant difference?

I think people should not think that much in terms of supply-demand for speculative phenomena. At least not in a macroeconomic sense.

The value of a coin is determined by the expectation of selling it in the future at a higher price, nothing else.

These expectations are affected by the supply level only via auction dynamics: lots of newly mined coins in a short timeframe tend to result in big sell orders and downward pressure on the price.

That's an effective but also very limited way to impact the price. For instance the increase in supply of a non-speculative asset reduces its marginal utility: the more oil available the less useful it becomes.

But there's no theoretical limit to the price of a coin and therefore no limit to the return one can expect!

As it's all about auction dynamics, a highly inflationary speculative asset with well-timed and strategic increase in supply could very well defy any macro supply-demand logic.

Well I wish he expanded a bit more on the “Just do X, that’s so much simpler” section.

He's on AWS, he could have gotten all those benefits with ElasticBeanstalk or ECS even. Plus no yaml files but actual IaaS (terraform etc.) and much better integrations with the other AWS services.

I'd personally still use EC2 for a blog, but if you're looking for convenience/battery-included type of thing...yes I'd argue you're better off just doing X instead.

When you can create money backed by nothing

Isn't that what money really is about?

Debt and money go together. When you take on debt your creditor can decide to sell his credit to somebody else, de facto using your promise to repay your debt as a currency.

Isn't your promise...just a promise? Yet you can buy stuff with it?

The central thesis of value investing is: if a company is going to close shop, the scrap value of that company will be higher than the market cap of that company. There's a lot more to it, but that's the gist of it.

There's enough "more to it", to make this statement incorrect.

You're describing Graham's style of value investing, Buffet's style is significantly different.

In fact the term "value investing" has now a very loose definition, since so many different styles are grouped under this term.

I'd say this is the most appropriate definition: "Investment is most intelligent when it is most businesslike" (straight from the Intelligent Investor).

In other terms, any type of investing that's not based on speculation is value investing.

They are certainly very successful ideas in terms of outcomes. Apple went up more than 3X since he bought it in 2016.

I'd say he broke his own game of investing, rather than the game in general.

Given what he thinks he understand there's not enough opportunities to allocate this amount of capital.

I guess that's why he hired the two new managers. By the way they did come up with some of the best ideas in recent years like Apple and Snowflake.

A (big?) part of Databricks' success is the complicated mess that Spark is to run.

That sounds wonderful (really)

Was going to say. Most of the times all it takes is to have a proper data model.

For analytics I favour de-nomarlized schemas and, if necessary, nested fields. Queries are much easier to write (fewer joins), much faster, no need to incrementally materialize (sigh), fewer backfills and no messy field definitions.

What you often see instead is highly-normalized data models with an un-trackable amount of materialized views (usually on top each others) and some complicated tools/solutions to try to deal with all that mess. The cost of a bad design.

You're not the product, it's a wrong analogy. You simply pay hidden fees embedded in the slightly higher price you pay.

Precisely.

Problem with SAM is having to deal with one more tool.

But also, maybe even important, it makes things less transparent. I want to be able to refer to the original CloudFormation docs (even when using Terraform) without "translating" every time.