HNHacker News
TopNewBestAskShowJobs

kccqzy

19,975 karma · joined April 8, 2014

GitHub: https://github.com/kccqzy

I love Big Tech more than I love startups. Feel free to disagree with me on that.

submissionscomments
kccqzy··on A thousand fewer paths to root
LPEs are frequent enough that people no longer use low-privilege user accounts to sandbox potentially malicious software. They are only used to control access for well-behaved users without any intent to subvert the system, not attackers.

The difference here reminds me of the distinction between the regular Paxos algorithm and the Byzantine Paxos algorithm.

kccqzy··on A new Go API for JSON
Yeah that’s what I mean by strictly. Does your implementation take shortcuts or does it hold itself to a higher standard and not do that? Does it only use available memory as the ultimate limit?

A lot of specs contain these optional limitations. The C standard allowed implementations to truncate variable names to 31 characters if they have external linkage and 63 if internal linkage. Would you be happy if your implementation actually does that? Of course not. Neither LLVM nor GCC does that. That’s because people writing compilers also hold themselves to a higher standard. Now imagine this kind of attitude being pervasive.

kccqzy··on A new Go API for JSON
Some languages have a culture of doing things correctly and conforming strictly to the spec. These languages have indeed sorted it out years ago. Haskell for example. Any language that doesn’t have a special number type for JSON is a half-assed implementation, because 1e1000 is a perfectly valid JSON number that is beyond the typical IEEE754 double precision floating point numbers; does your language handle that correctly? Haskell does. This is but a small example of a strong culture bias towards correctness.
kccqzy··on Mac at any scale
People love to throw out these buzzwords describing their ideal “maximum security” setup suitable for “sensitive production workloads” without ever defining what the threat model is.

The most grating thing is requiring every workload to be in a dedicated VM. You are throwing so much performance away. What kind of threat model would justify this? Usually there’s no threat model analysis at all.

Banning opensshd is again mostly a fool’s errand. It’s much more productive to design a system that writes the authorized_keys file read by opensshd. For example you can design a system that requires peer approval before your own key gets written to authorized_keys.

kccqzy··on Safari introduces Adaptive Dark Mode in 27.3
A pet peeve of mine is websites thinking oh you prefer dark mode so let’s make the background pure black. I cannot take it because the contrast is too much: #fff on #000 is way more uncomfortable than #000 on #fff for me. There are certain developers who continue to think that dark mode is for conserving battery life with an OLED screen so pure black is preferable. I hate that so much that I have turned off dark mode globally. If websites can’t be bothered to implement an actually nice dark theme like Solarized Dark, I’d rather prefer the light mode.

Besides Solarized Dark, GitHub’s Soft Dark theme is quite well done, but unfortunately it’s not the default for a logged-out user with dark mode. So with dark mode I need to log in to GitHub to get acceptable colors.

Now that I think about it, I need a feature to limit the darkness of colors anywhere on the page, background or text except images. Remap #000 to #222 and #fff to #ddd.

kccqzy··on Safari introduces Adaptive Dark Mode in 27.3
Wikipedia allows you to write any custom CSS. In 2012/2013 when Facebook IPO’ed and was the cool thing I made a user CSS that made Wikipedia look like Facebook.
kccqzy··on Safari introduces Adaptive Dark Mode in 27.3
It’s not. The feature is implemented deep within WebKit so it makes sense to be on this blog.

> It’s precisely because Adaptive Dark Mode is happening deep within the WebKit browser engine that we are able to find creative ways to solve the most complex challenges, apply the transformations with great speed and power efficiency

kccqzy··on Safari introduces Adaptive Dark Mode in 27.3
Browsers lie all the time. Browsers also prevent sites from getting the color of a link because visited links can be shown in a different color and this can be used by the site to tell if a user has ever visited a link.

Don’t you love it when browsers lie to websites to enhance the user’s privacy.

kccqzy··on Fujifilm to boost tape production 5x as AI fuels archiving demand
Historians actually would love this. Currently many historians try to reconstruct how life was like centuries ago from accidentally preserved artifacts. Future historians would have no trouble understanding how our daily life is like.
kccqzy··on Mac at any scale
Amazing article. It just goes to show how many problems they are solving just to work around bugs and limitations within Apple itself. I’d wager a guess that if Apple made a new Xserve with the same performance as a Mac mini but twice the price people would still buy them, because figuring out data center operations is expensive.
kccqzy··on Google will turn off search for schools that reject AI
> … and Google Search. Access to these products will be entirely off for K-12 student accounts until school administrators explicitly turn them on.

So yeah the blog post says Google Search itself will be entirely off until being turned on. There’s nothing in this blog post to indicate whether Google Search can be turned on without also turning on Gemini. Although I hate the NYT, the NYT is quoting someone from Google saying there won’t be separate controls for Gemini and Google Search.

kccqzy··on Astra can read and write weights to improve Qwen
Amazing experiment. Indeed, with the benefit of hindsight, we can say that breakout is a game that is probably better suited to deterministic code (it's not like Go where an AlphaZero style training is essential), and Astra is probably translating this internal representation of code into the shape of linear algebra and model weights.

I should note that the original authors did not allow Astra to write any significant code to construct the weights. I think this is unfair. If you allow strided assignments to numpy arrays, you should probably also allow the use of functions like np.linspace, np.repeat, and np.tile.

Oh and Astra wrote this (https://github.com/RikiyaT/lm_understand-logs/blob/4d09ba6d6...):

    W1[63:120:6, 5] = 10
    W1[123:180:6, 5] = 20
    W1[183:240:6, 5] = 30
    W1[243:300:6, 5] = 40
    W1[303:360:6, 5] = 50
    W1[363:420:6, 5] = 60
    W1[423:480:6, 5] = 70
    W1[483:540:6, 5] = 70
    W1[543:600:6, 5] = 90
I think the second 70 there is probably a typo; it looks like it should be 80 instead.
kccqzy··on Why are Soviet math textbooks so hardcore in comparison to US textbooks? (2017)
My first undergraduate calculus textbook was Advanced Calculus by Gerald Folland. It’s not colorful. It’s serious. I remember it to be so concise that I wished it gave more examples.

Maybe you are thinking of high school or middle school textbooks?

kccqzy··on Intel stock collapse worsens as $3.5T chip rout spreads
In the worst case when the free market deems Intel worthless, the government will nationalize Intel at a cheap nominal price. Shareholders are wiped out anyways.
kccqzy··on BYD sold more EVs last month than the US did all quarter
If not, just don’t buy it. Let the free market sort it out. If the new part is proven to be good, the market will price these cars much higher, and sellers will perform a preemptive replacement to get higher resale value.
kccqzy··on BYD sold more EVs last month than the US did all quarter
It’s simple; if the used model has undergone an ICCU replacement in 2026 it will be better. It has a new part number, 36401-1XAD6.
kccqzy··on Anthropic's growth just flattened
It may have something to do with the summer intern season ending. I know one of the interns hired by my employer used up the allotted AI spend by doing “fun” things unrelated to the internship project, like making Claude write an entire SQL-to-Pandas transpiler in 10,000 lines of Python. I can’t blame them for using it extensively when given access to a generous limit on token usage.
kccqzy··on BYD sold more EVs last month than the US did all quarter
You are repeating years-old news by now. It’s true that the 2022 through 2025 model years had the ICCU problem, but they improved the ICCU design in 2026, with both better hardware (heavier duty power electronics) and better software (slow start for 12V charging).

Of course time will tell whether this brings failure rate down to be negligible. But I feel pretty confident.

kccqzy··on Astra can read and write weights to improve Qwen
The headline makes it sound very amazing, but reading the reasoning loops it didn’t strike me as outlandish. It feels no different from giving a human expert a printout of assembly code and figure out what it does. (The human expert can do it without tools too, but they probably wouldn’t do it because it’s tedious.)
kccqzy··on Why are Soviet math textbooks so hardcore in comparison to US textbooks? (2017)
Textbooks don’t need a lot of revisions though. Basic math has not really changed that much. One of my favorite textbooks I used as a kid was written ten years before I was born.
kccqzy··on Why are Soviet math textbooks so hardcore in comparison to US textbooks? (2017)
The U.S. textbooks are so much better than textbooks in many other countries. They are thick and heavy, but then that is because they tend to explain the same concept in multiple different ways so students with different thinking modes can find their own explanation that clicks.

Furthermore, having good textbooks papers over the fact that in the U.S. the qualification and ability of teachers vary so much across poorer and richer school districts. A smart student in a poor school district can use the textbooks to a greater advantage.

kccqzy··on WTF Ubuntu?
Suppose you are an enterprise paying large amount of money to NVidia to buy their AI hardware to run AI workloads locally. Are you going to say we will refuse the default OS (https://docs.nvidia.com/dgx/dgx-os-7-user-guide/introduction...) on such hardware and lose hardware support, or are you going to stick with the supported OS for your hardware?

Enterprises care far more about vendor support than the typical Linux hacker at home.

kccqzy··on Mathematics has changed. Let's act like it
> To me, someone who chooses a worthwhile problem, obtains a proof completely using AI, understands it thoroughly, and uses it to develop interesting further questions or directions has made a substantial mathematical contribution. I would give that person substantial credit even if they supplied no new ingredient in the original proof.

Yup. This is what Terence Tao has been saying about digestion of results for several months now. Mathematics has become a field where results themselves matter less, the social aspect of attracting others to understand your results and explore new problems has become more important.

In a sense it is a little bit like software engineering, where writing code can sometimes be delegated to AI, and is arguably less important than understanding what the customer really wants from this software.

kccqzy··on Bidirectional charging during grid outages
You don’t need to run your 5+ ton HVAC full time. Get an EV and run the AC there (no more than 3kW). Sleep in the EV. A typical 80kWh EV battery will last several days. And this is not just theoretical; plenty of people have done this.

For Christ’s sake a natural disaster just struck and you want to live like nothing happened? The entitlement is unbelievable.

kccqzy··on Domain hacks with unusual Unicode characters (2018)
My browser refuses to open this link. It couldn’t find the server.
kccqzy··on AI, Lean, and the Future of Formal Verification, with Lean Creator Leo de Moura
Property based testing is great because it’s implemented as a library in many languages. You don’t have to rewrite your codebase in Lean.
kccqzy··on AI isn't replacing SREs, it's deskilling them
Every incident having a root cause is required. If your organization doesn’t do this, it basically means they don’t value reliability at all. The best SREs quit. Well either that or the SREs have the ability to transfer oncall responsibility to SWEs. Whether or not that can be fixed at the origin is an entirely different matter. Bug-free code doesn’t really matter: the best incidents are the kind that reveals hitherto unknown bugs.
kccqzy··on US tech stocks see sharp falls over AI cost fears
Google’s AI mode is roughly the same as using the free ChatGPT web app (both without logging in). Of course they use a lower-intelligence model for free users. They are certainly worse than real AI subscriptions. I do the same thing you did: only use it for basics and low-stake things, and it’s good enough for basics.
kccqzy··on I didn't know I was lacking in Vitamin D until I collapsed at work
PSA that deficiency in Vitamin D also leads to cognitive decline including dementia. https://pmc.ncbi.nlm.nih.gov/articles/PMC6132681/

I find the risk-award very clear: if you are not deficient, taking normal* amount of Vitamin D supplement has no adverse effects; but if you are deficient, the risk is quite high.

*: The normal amount is 600 IU. The recommended upper limit is 4000 IU. The gap is wide.

kccqzy··on AI isn't replacing SREs, it's deskilling them
I stopped reading halfway because it doesn’t describe competent SREs in an organization where the culture values reliability.

> Alert noise reduction and intelligent grouping, runbook execution for known issues, log pattern detection and anomaly flagging, and basic root-cause suggestions from historical incident data, and auto-remediation for well-understood failure modes like restarting a crashed pod or scaling up a service that’s running hot, are all fairly automated today.

Let’s stop here and ask, why do we not fix the underlying issue? Alert noise? Design better alerts. Runbook execution? Make it a cronjob. Known issues? Why aren’t people fixing these known issues? Root cause suggestion from historical data? Sorry if root cause A is in the historical data you expect problem A is fully solved so it wouldn’t appear again in the next incident. Restarting things? Why doesn’t your health monitor restart things automatically?

SREs also need to climb up the abstraction ladder. Instead of using AI to assist with grunt work, use AI or human thinking to build more reliable systems with fewer incidents. Every incident should be novel.

← PreviousPage 3 of 34Next →