923 karma · joined July 6, 2018
Take, for example, PHP… look at the features released in the last 6 or so years, starting with PHP 7, and how mature the language has become.
With the advance of AI-assisted programming, I feel like Python is always a bad choice.
[1] https://github.com/cdlewis/snowboardkids2-decomp/tree/main
I’m doubtful this is a meaningful result. Kernel contains a lot of legacy code and generalizations to support different hardware etc.; removing that would result in a speedup. Next are all the mitigations for hardware vulnerabilities and attacks. If removed would give a nice speedup as well at the cost of security. And then finally, just specializing the Kernel in whatever the benchmark is measuring, making it useless as a general piece of software would also make it fast.
I’m learning Japanese, and I’ve started learning Chinese characters, both their meanings and how to read them. Reading them feels different than English... I have a hypothesis that our brains work differently when processing symbols that encode meanings as opposed to just sounds. English requires an extra step, where characters are translated into sounds and then into words.
With Chinese characters, you are immediately looking at the meaning; you don’t need translation into sounds. This feels like a more efficient process cognitively to me, even though I have to memorize to recognize more characters.
They all roughly produce junior developer-level code, continue to have mental breakdowns in their “thinking” stage, occasionally hallucinate things, delete pieces of code/docs they don’t understand or don’t like, use 1.5 times the necessary words to explain things when generating docs and so on.
I'm now testing "avoid sycophancy, keep details short and focus on the facts" in my AGENTS.md files.
Somehow, employers of these establishments convinced the staff that it's the customer’s fault that their wages are inadequate and that they should go after the customers to get the difference. I would much rather pay a higher price and not hear anything about the tips.
> A diff can show what changed in the artifact, but it cannot explain which requirement demanded the change, which constraint shaped it, or which tradeoff caused one structure to be chosen over another.
That's not true... diffs would be traceable to commits and PRs, which in turn are traceable to the tickets. And then there would be tests. With all that, it would be trivial to understand the whys.
You need both the business requirements and the code. One can't replace the other. If you attempt to describe technical requirements precisely, you'll inevitably end up writing the code, at very least, a pseudocode.
As for regenerating the deleted code out of business requirements alone, that won't work cleanly most of the time. Because there are technical constraints and technical debt.
I currently own a Lenovo Legion laptop. Still, a very powerful machine, but the screen now has a spot in the middle with multiple dead pixels, the topcoat on the trackpad is peeling off, and the main body has spots where palms rest. I'd happily buy replacement parts and install them, but I can't.
If they were to measure a person who performs mental arithmetic on a daily basis, I'd expect his brain activity and oxygen consumption to be lower than those of a person who never does it. How much difference would that make?
Even if this was an accident, isn't it theoretically possible for one of the trustees to intentionally not provide the key to trigger the re-election? There's no guarantee that the people will vote the same. I see this as a kind of vulnerability.
Essentially, predicting stock movement becomes predicting the sentiment of the people. Instead of "the financial report means the company is doing X", it becomes "if people were to see this financial report, they would react by X".
And this whole thing feels like a big Ponzi scheme. Everyone keeps repeating that you need to invest your money, and that's what essentially makes the market long-term bullish.
Except I wanted more security and multiple users. Instead of using the default admin user, I created one user for each person. Done in the "_users" database. Then create one database for each person. Assign each user as a "Member" to their respective database, not admin. Now each person has their own credentials that can access only their database.
Nextcloud can't even get Notes done right. I lost the entire contents of the note randomly not long ago. And the mobile Note app refuses to load the editor sometimes.
That being said, most of the time, Nextcloud works ok. I don't want to replace Nextcloud with another jack of all trades, master of none. Instead, I'm slowly migrating to good alternatives that do one thing well: Immich for photos, Obsidian for notes.
And that's why I will never buy any IoT devices that require an internet connection to work. Only IoT devices in my house are those that connect to my own server and never see the light of the internet.
Google flagged my domain as dangerous once. I do host Jellyfin, Immish, and NextCloud. I run an IP whitelist on the router. All packets from IPs that are not whitelisted are dropped. There are no links to my domain on the internet. At any time, there are 2-3 IPs belonging to me and my family that can load the website. I never whitelisted Google IPs.
How on earth did Google manage to determine that my domain is dangerous?
But I do have Zigbee sensors and switches, all of which connect to my home server and Home Assistant. None of them see the internet. But Home Assistant is accessible from the internet through a reverse proxy from whitelisted IPs.
https://github.com/PrivSec-dev/banking-apps-compat-report
https://privsec.dev/posts/android/banking-applications-compa...
> The certificates provide no security: The way you verify your identity to Let's Encrypt... you place a file somewhere on your website, and they access > Automatic renewal is insecure: certbot ... downloads a bunch of untrusted data from the web, and then feeds that data into your web server, all as root...
False, this is NOT the only way. You can do it by setting a TXT DNS record. No files involved. Your server communicates with the registrar through API over an encrypted connection.
> Manual renewal isn't free...
You're not supposed to manually review Let's Encrypt in the first place. The whole point is that you set up cron job once and forget about it.
> HTTPS is a trap: Once you've moved your websites to HTTPS, there's no going back to plain HTTP...
Not true. You can run HTTP version of your website in parallel if you so desire.
> Let's Encrypt is founded on the benevolence of scoundrels: Let's Encrypt isn't free to run, either. Their 2019 operating budget is 3.6 million U.S. dollars. Most of that is donated by… guess who? _Your competitors_[1].
EFF and Mozilla foundation are not my competitors LOL.
[1] https://www.abetterinternet.org/sponsors/
>It's bad engineering: When you install a certificate with a three-month expiration...
For a business, there's no incentive anymore to choose Germany over other EU countries. Cheap energy is gone. Germany is literally collapsing under its own weight: endless bureaucratic structures that just keep on growing.
I don't see how some kind of big breakthrough is going to happen with the current model designs. The superintelligence, if it will ever be created, will require a new breakthrough in model architecture. We've pretty much hit the limit of what is possible with current LLMs. The improvements are marginal at this point.
Secondly, hypothetically, the US achieves superintelligence, what is stopping China from doing the same in a month or two, for example?
Even if China achieves a big breakthrough first, it may benefit the rest of the world.