HNHacker News
TopNewBestAskShowJobs

kjksf

6,940 karma · joined September 5, 2009

My website is https://blog.kowalczyk.info

Working on https://filerion.com/, a web based file manager (s3, Dropbox and more).

I created https://www.sumatrapdfreader.org

submissionscomments
kjksf··on The withering dream of a cheap American electric car
Inflation is a shit metric because it's easily manipulated.

The cost of a GB of hard drive is failing spectacularly. The price of health care went up much more than the price of eggs. So what is the "real" inflation?

Government gets to pick what they use to define inflation so they can manipulate "inflation" numbers. And manipulate they do.

What you should look at is money printing: how much money did the government print. This is about 8% yearly for US.

This money debasement is eventually reflected in prices.

Some things get cheaper, because we can produce them more efficiently (like hard drives). Some things get even more expensive than 8% because we produce them less efficiently (health care insurance or college diplomas).

So to answer your question: cars costs more mostly because the government prints money, which devalues your dollars and car makers are not getting more efficient at making cars to counter currency debasement.

kjksf··on The withering dream of a cheap American electric car
It's not an observation because it's not something that you've observed.

Here's the truth based on an observation: Tesla's battery capacity degrades 12% after 200k miles. Source: https://insideevs.com/news/664106/tesla-battery-capacity-deg...

200k miles is effectively the lifetime of a car. Average US person drives 10k miles so that's 20 years of driving.

Tesla's warranty "guarantee at least 70 percent retention of battery capacity over 8 years and 100,000 miles or more". Source: https://www.motortrend.com/features/tesla-battery-warranty/

And latest chemistries are even better. In 2020 Jeff Dahn (who leads battery research group in Canada funded by Tesla) published a paper about million mile battery. Source: https://www.electrochem.org/dahn-unveils-million-mile-batter...

Since Tesla funds Dahn's research, they get the IP. This is just in the lab but those advancements are trickling, over time, to Tesla's battery making (and not just Tesla: every battery maker does research to make batteries cheaper and last longer).

kjksf··on The withering dream of a cheap American electric car
Tesla cars sold in US are the most American build cars. See https://www.carpro.com/blog/most-american-made-vehicle-the-t...

What it means that if you count what percentage of car parts are made in America, Tesla has higher percentage that other brands, even those you might consider "more" American, like GM or Ford. All Teslas sold in America are assembled in America (California or Texas). Ford Mustang, for example, is assembled in Mexico.

As far as I know Tesla never sold Chinese built cars in US. They used Chinese manufacturing (CATL) and Korean batteries in some model, but also manufactured batteries in US (Nevada, with Panasonic) and are expanding battery production in US with 4680 (used in Cybertruck).

Cheaper Teslas are coming to US but they'll be manufactured in US (Texas). Tesla told us that they'll start making a cheaper model sometime in 2025.

kjksf··on Looking for a Job Is Tough
Name a prominent open source project created be and led by a woman.

Who created and contributes to linux kernel, python, c, c++, go, python, ruby, ruby on rails, php, wordpress, ghost, SumatraPDF, zig, oding, nim, nodejs, deno, bun.

I could just keep listing major open source projects because I literally cannot think of a single one created and led by a woman.

And if you find one, it still doesn't negate the 100 to 1 ratio.

Open source, unlike jobs, are pure meritocracy. A woman can create a GitHub account and start coding just as easily as a man. There are no gatekeepers and open source contributors / maintainers are abused by random people as a matter of course.

To me it's reality. To you, somehow, saying that out loud is bias.

And to be clear: I don't think there's anything preventing women from learning to code and contributing at a high level and I've known a few that do. But for some reason they overwhelmingly don't. The stats are brutal.

kjksf··on New York Times Tech Guild goes on strike
If you ever went through interview loop at Google or a similar company, I doubt you would call those companies "lazy" wrt. hiring.

An interview is at least 4 people, each grilling you for an hour, asking hard questions.

Their hiring bar is high and they optimize for avoiding bad hires (which of course is pissing off the commenters who want to be hired and therefore would prefer lower hiring standard).

In Europe they make it harder to fire people and guess what happened?

First, companies have probatory period (2-6 months, depending on the country) where you're hired but can be fired at will. This is to minimize chances of being stuck with a poor performer.

Second, EU economy is about the size of US and China but software industry (and the tax / employment riches associated with it) is largely in US Chine. Might be a coincidence but I think there's causality between over-regulation and stagnation of the economy.

kjksf··on New York Times Tech Guild goes on strike
That might by what YOU want or what you hallucinate the demand is but most reasonable interpretation of what we know is that they in fact want to prevent being fired for low performance.

if you can be fired "only for misconduct" and low performance doesn't count as misconduct means that you cannot be fired for low performance.

Granted, the actual demand might be more nuanced but going only by what was reported, they don't want to be fired for low performance.

kjksf··on If you need the money, don't take the job
In the context of the article, fixed pricing means pricing per project.

Charging per month of work is not that.

kjksf··on Astral: Next-Gen Python Tooling
Then you'll have a very good, open-source, MIT licensed tool financed by silly VCs.

If it gets abandoned the community will be able to take over its maintenance or fork you.

Tails, you win, heads, you win.

https://github.com/astral-sh/ruff/blob/main/LICENSE

kjksf··on 10 Years of HexaPDF
You should be able to see when someone is gaslighting you. They charge because they want and can.

The above justification might apply to some of their standards but certainly not to PDF.

PDF was developed by Adobe. They make lots money selling PDF tools and licensing PDF software.

PDF didn't have to be a standard. It could be a proprietary format.

Adobe wanted the adoption of PDF so they made PDF a standards so that they can sell more PDF tools.

It didn't have to an ISO standard. Adobe could have released it as an open spec with some sort of permissive license.

Adobe wanted it to be ISO standard to piggy back on "respectability" of International Standards Organization, to buy the impression that it's an open standard.

But ISO makes money by selling standards so we ended up in this situation when the spec for "open" standard that Adobe wants to be adapted costs $100+ to buy.

I guess Adobe at some point figured out this is stupid and made the spec available for free.

kjksf··on The EU Throws a Hand Grenade on Software Liability
I wonder how do you expect a high school kid or collage student to afford insurance and get "some sort of license to practice"?

Say goodbye to Linux - released by college student.

Say goodbye to Winamp - released by a high schooler.

Say goodbye to GTK+ and GIMP - released by 2 college students.

Say goodbye to web browsers - why would Tim Barners-Lee or his employer CERN release a free product not related to their core mission if they could get sued by thousands of people?

kjksf··on Show HN: Go Plan9 Memo
It's more "reused" because Plan9 compilers were designed / co-implemented by Pike / Thomson, 2 out of 3 original designers of Go.

For those interested, here's Thomson's paper about Plan9 C compilers: https://9p.io/sys/doc/compiler.html and https://doc.cat-v.org/bell_labs/new_c_compilers/

kjksf··on Type 2 diabetes: New treatment eliminates insulin for 86% of patients
A big part of the problem is that if you go to a doctor with diabetes, 90%+ he will tell you to use insulin. He won't tell you to loose weight and go on keto diet.

Still popular opinion is that eating meat and fat is bad for you (heart attack) but no many understand that eating sugar and carbs is a highway to diabetes).

In fact many people who go on keto and reverse diabetes report that doctors instead of congratulating them and telling other patients to do the same, tell them that keto diet (i.e. eating lots of meat) will give them heart attacks.

Most people don't know how bad sugar and carbs are because no one tells them.

kjksf··on Show HN: Sava OS – A desktop interface for your web browser
If you asked me to login with Google and got my email and other info, then don't ask me to also pick a password. Bait & switch and I won't stand for it.
kjksf··on Pivotal Tracker will shut down
Free business idea: clone Pivotal Tracker as a solo dev / small team.

People often ask: how do I find business ideas?

Well, here you go: many people publicly saying how they love a product that is going away.

This is a validated product: people were paying for it. Apparently quite a lot of people. It doesn't get better than this.

All you have to do is to clone the product. You can literally market it as a Pivotal Tracker clone. It's not like VMWare will care.

You can research companies currently using Pivotal Tracker and build a database for cold calling / e-mailing when you have the product.

It's also a product that is doable as a single person or very small team. With modern technologies (React or Svelte, hosted databases etc.) it's relatively simple to clone.

Staying small is important: those businesses topple over when revenues don't justify expenses, especially if VC funding is involved and VCs are pressuring for going big or going bust. Or when a profitable product is acquired with the hopes of growing the profits but they don't grow.

Stay small to keep expenses in check and you can build a profitable company.

This is a bootstrappable business: a $100/mo Hetzner box, backend in efficient language (Go, C#), front-end in Svelte or React and you can serve lots of customers. The rest is your time and hustle.

kjksf··on How to succeed in MrBeast production (Leaked PDF)
It's more about willing to fire below-A players quickly rather than having a perfect hiring filter that only lets A players in.

Looking back at 7 companies I worked at: they all had a tough hiring filter to get in. But most of them also had not that great people that they were not firing.

Firing people is hard even when you know you should do it. You have to be a heartless bastard to not have a problem firing people.

It's even worse when the company gets so big that a game of building empires starts in which case managers have an incentive to grow headcount to grow power, even if that headcount isn't very good.

The document even talks about what MrBeast considers a B-player.

Made a mistake once? That's fine. Fuck ups are a price of ambition.

Made the same mistake twice? Need to be told the same thing multiple times? Not an A player so fired.

kjksf··on Human drivers keep rear-ending Waymos
Consider famous autonomous Uber crash that killed a person illegally crossing the road at night.

The right thing to do for the autonomous software would be to break hard to avoid killing that person.

If there was a car behind that Uber car, it would likely rear end the Uber. That would be preferable to killing the human. The software would be right. Unpredictable but right.

If the read ending happens because Waymo erroneously did phantom break (i.e. it did break because software was confused), then software would be wrong.

Without analyzing rear ending accidents to decide if the software was right or wrong, we can't say if Waymo did the right thing or not.

In human-to-human crashes, rear-ender is presumed guilty.

kjksf··on CS Career fair cancelled at community college because no companies reached out
The "washout" would work equally well if they were teaching non-BS.

And what do we teach people for 12 years before college? Is college the first time people learn how to learn?

kjksf··on Don't defer Close() on writable files (2017)
It's a trivial function: https://cs.opensource.google/go/go/+/refs/tags/go1.23.1:src/...

You could write your own errorsJoin() and change Error() method to suit your needs.

But really in this particular scenario you would be better served by something like:

    func errorsConcat(err1 error, err2 error) error {
      if (err1) {
        return err1;
      }
      return err2;
    }
And then do: err = errorsConcat(err, f.Close())

In the scenario described in this article, errors.Join() would most often reduce to that (in terms of what Error() string would produce).

kjksf··on Judge stops FTC from enforcing ban on non-compete agreements
Dear god.

Do you understand what you're advocating for? You're advocating for literal lawlessness. For anarchy. For giving the government absolute power.

Given that government abuses non-absolute power you can bet that when they get absolute power it'll take about a second to turn any country into Putin's Russia.

kjksf··on Judge stops FTC from enforcing ban on non-compete agreements
Well, yes, because separation of powers.

Congress makes the laws.

Government (bureaucrats) enforces the law.

Judges decide if government enforces the law correctly.

The important part is not how bureaucrats or judges get their job but that bureaucrats were de facto creating laws exceeding their authority.

Because the authority to create laws belongs to congress.

It's not a perfect system but it would be even worse without separation of powers.

kjksf··on Blood puddles, mold, tainted meat, bugs: Boar's Head inspections are horrifying
> According to the CDC, one significant outbreak in 2020 was linked to onions, resulting in around 1,127 reported cases of Salmonella across 48 states.

Vegetables can also be contaminated with bad bacteria and they do get contaminated in practice.

There are plenty outbreaks caused by contaminated salads and other vegetables.

kjksf··on Man vs. AI: optimizing pdf.js JavaScript function
I decided to test if AI (Cursor + Claude 3.5 Sonnet) can optimize JavaScript code for me.

It can and the above recaps my journey.

kjksf··on The 4-chan Go programmer
Assuming latest Go 1.13 I would write an iterator and used goroutines internally.

The caller would do:

    for f := range asyncDirIter(dir) {
    }
Better than exposing channel.

But my first question would be: is it really necessary? Are you really scanning such large directories to make async dir traversal beneficial?

I actually did that once but that that was for a program that scanned the whole drive. I wouldn't do it for scanning a local drive with 100 files.

Finally, you re-defined "traversing a tree" into "traversing a filesystem".

I assume that the post you're responding to was talking about traversing a tree structure in memory. In that context using goroutines is an overkill. Harder to implement, harder to use and slower.

kjksf··on Ask HN: What are you working on (August 2024)?
I'm working on Edna - a note taking app for developers and power users.

https://edna.arslexis.io/help

kjksf··on Go is my hammer, and everything is a nail
Maybe, but 80% of those are complains of some sort.

That's the opposite of hype.

kjksf··on Go is my hammer, and everything is a nail
So you're ok with Python, which doesn't have sum types or null safety, 2 out of 3 things you say are missing in Go.

In fact, Python didn't have any compile-type type until recently and compared to Go is slow and bloated.

So maybe the issue is not Go's lacking some features but something entirely different.

kjksf··on Bitcoin Investors Buy Puts Aggressively
Did you know it before reading the article?

If not, then yes, it's news. As in: new information.

What you're probably trying to say is that it's not valuable information.

Maybe it isn't.

However the belief in automated trading being some kind of God-like force in trading that vanquishes all other opponents, while common, seems to me very unfounded.

If it was possible to build automated trading to feed off a simple signal like put / call ration then everyone would do it and then no-one would have any edge.

The thing about markets is that there are many participants and they believe different things.

Some people believe that this particular signal predicts something about future prices and some believe it doesn't.

You make money if your believes prove correct. You make most money if your believes prove correct while most other market participants believe the opposite.

For example, if you believed in Bitcoin 5 years ago, you made a killing.

If you believe in Bitcoin today, you'll make money if you're right about demand for Bitcoin going up. And if you'll make money, then you'll probably make a killing because even today most people do not believe in Bitcoin.

kjksf··on Bitcoin Investors Buy Puts Aggressively
I can "divide" a single $100 bill into 100 $1 bills.

By your "logic" $1 is just as valuable as $100.

kjksf··on Bitcoin Investors Buy Puts Aggressively
Bitcoin is deflationary in the sense that the supply is fixed at 21 million. And because people also loose access to their Bitcoin wallet, the actual supply will slowly decrease.

That's just a fact of how technology works. It's better phrased as "supply of Bitcoin is fixed".

As to your other arguments: I don't really get the connecting logic or even the point your making.

What is the "spread" that will stop new people from buying Bitcoin?

Can you explain the "Tom got very rich on Bitcoin and therefore I won't buy it today even though the price goes up because reasons..." logic?

Also what does reward halving have to do with the price of Bitcoin? The mining business will eventually end but as long as Bitcoin price goes 2x by each halving, the mining is is just as profitable.

Of course Bitcoin price can collapse. It already did a couple of times. But it also recovered and then some each time.

Given that prices of anything are driven by supply and demand, supply is fixed, the only question worth asking about Bitcoin is: will the demand go up? If demand goes up, the price goes up. And vice versa.

kjksf··on Bitcoin Investors Buy Puts Aggressively
So just two days ago the market crashed because apparently a yen carry trade unwound.

Japan had a 0% interest rate for very long time. That made yen weak against other currencies. So people were borrowing Yen and investing in better performing assets. That's leverage.

Japan raised interest rate every so slightly which made Yen jump comparatively to other currencies which meant all those people who over-leveraged had paper losses, margin calls and had to emergency sell other assets hence the market crash.

There's a lot of "gambling" in financial markets. There's a lot of "gambling" in business.

You can't eat gold. Buying gold is a gamble that the value of gold will go up in the future.

You can't eat a Picasso. Buying a Picasso is a gamble that the value of paintings by dead people will go up in the future.

A business loan is a gamble that you can build a business that generates a profit.

A mortgage is a gamble that the value of the house will go up and value of the dollar will go down. In deflation people with mortgages will be screwed.

I could go on. I don't get why people think that singling out Bitcoin as speculative asset is some sort of "end of discussion" revelation. Ultimately, even the safe, cash flow generating S&P 500 can permanently crash given a black swan event.

As someone in finance you should know that the fundamental law driving prices of assets is supply and demand.

The supply of Bitcoin is fixed. There will never be more land in Manhattan and there will never be more than 21 million Bitcoins.

So the only question you should be asking yourself is: will there be more demand for Bitcoin in the future?

And as of today the answer is: hell yeah.

First US Bitcoin ETFs were approved this year. They generated inflows of over 30 billion USD to Bitcoin market cap. They were the most successful ETFs EVER.

That was followed by ETFs in other countries and there are even more.

And we're just at the beginning. Pension funds, banks, hedge funds, companies and other large holders of financial assets only begun converting some of those assets into Bitcoin. Barely wetting their beaks. Sips will follow, heavy drinking will follow.

Some countries began buying Bitcoin as banking reserves, similar to how they buy gold. There are talks about US doing so.

All this might change so in that sense buying Bitcoin is a gamble but for the next few years the prognosis is clear: up and to the right.

← PreviousPage 7 of 34Next →