HN user
I usually don't know what I'm talking about.
And -march=native or at least -march=x86-64-v3 or similar, alternatively identifying relevant functions and manually invoking FMV and uarch specialization via target_clones. Plus non-integer code can generally not be autovectorized in normal-math mode since FP is non-commutative.
Spec-driven development is sold on the idea that you regenerate modules or entire services when you change the spec. It's a really dumb idea but was briefly popular in enterprise, probably still is in some companies.
Not user code, no. Someone eventually has to, but virtually every ORM under the sun allows you to construct dynamic queries without having to concatenate strings yourself or resort to string interpolation.
Most problems should be divide-and-conquerable into chunks that 300 vs 400k context is basically never an issue. A coding agent isn't an infinite chat.
You can even see the clear download slope starting in 2016 and (sans COVID) not changing slope substantially until about 2023. This part was not caused by AI at all.
This situation has.. no precedent as far as I can tell..
No, this has been going on for years. Vendors have been pushing malicious software through the Windows Update automatic driver installation since forever. MSI and Nahimic/A-Volute (this has watchdog daemon to instantly reinstall it as well as the main app protecting the daemon), the ASUS Armory Crate bullshit, the Lenovo garbage, which initially they only put into their own images, but then started force-installing via Windows Update, Gigabyte, ... the list is really long.
If you have to use Windows, you really absolutely should disable driver installation through Windows Update.
That doesn't block writers (when the writer uses WAL)
Neither does VACUUM INTO or ".backup" (which uses the backup API) or sqlite3_rsync or litestream.
It randomly and frequently turns off, then the connected services of course no longer work, then you turn it on manually, they still don’t work, then you hard kill safari a few times and/or toggle it on and off and it eventually works. On Linux or Windows the client is rock-solid however.
Meanwhile with wireguard: It just works. Every time. Every where. Unless someone blocks UDP.
Tailscale on iOS has always been highly unreliable for me, while the regular Wireguard app just works.
libssh, libssh2. These are totally independent and unrelated code bases, libssh is maintained by Red Hat mainly for ansible and some other tools, libssh2 was created for curl. libssh2 is client-only, libssh can also be used to implement servers.
The SQLite documentation doesn't say that. WITH and WITHOUT ROWID tables are also different data structures (ordinary tables are B+ trees, without rowid B*). In particular, without rowid tends to be detrimental for tables consisting of wide rows, while being an advantage for narrow rows with a non-integer key.
This document describes the support for SQL foreign key constraints introduced in SQLite version 3.6.19 (2009-10-14).
If every time the SQLite team added a better default for something they changed it, we'd be at SQLite 11 by now and every application would contain at least six incompatible versions of SQLite.
Steve declared thermonuclear war on Google because Android re-skinned to use BUTTONS.
Was there ever a point in time where Google was not the default search engine on iOS?
sqli is easily and fully mitigated and has generally been a non-issue for any half-serious project, especially if you use any kind of SAST. Your link actually subsumes any type of injection, not just sqli. Some of them are marginally harder to fix than sqli, most aren't.
In contrast, we don't know how to solve prompt injection.
Console vs PC is a red herring since UE games still stutter on consoles, people just notice that less because most games run at 30 fps anyway. You can read Unreal's own blog posts on this and they'll actually explain that this is mostly down to their material system (and game logic/scripts reaching into it) being designed to create nearly infinite shader variations on the fly in response to arbitrary world/game states. This design choice separates engines which have shader stutters from those which do not.
https://www.unrealengine.com/tech-blog/game-engines-and-shad... https://medium.com/@GroundZer0/what-unreal-doesnt-tell-you-a... https://therealmjp.github.io/posts/shader-permutations-part1...
UE is also just not a good engine. If you reduce "what is a 3D game engine" really hard it is "framework for running an input-process-output loop while consistently rendering audio and video". UE fails at the latter (factually unfixable shadder stuttering issues). A pragmatist might consider the taxonomy and say UE is therefore not a game engine. Unreal seems to be a great framework for assembling and generating content, though. Maybe Unreal Free was the metaverse all along?
Meanwhile idTech certainly had issues, mostly in regards to dynamic levels, especially in older iterations, but "consistently rendering audio and video" was certainly never among them. It is well above the industry mean when it comes to the core of what a game engine is.
Most of the time people are not asking for domains which don't resolve. Also, think about what DNSSEC is actually having to do when signing an NXDOMAIN. It needs to prove a negative with offline signing keys, DNSSEC does this by basically making a linked list of each zone and signing the links.
It leveled a huge stretch of coast, but didn't make it uninhabitable.
This is not actually true, because it deleted around 120 km² of habitable land (https://www.forbes.com/sites/oshadavidson/2011/04/29/japan-e...). Engineering just to reclaim that land continues to this day and has already cost trillions of Yen. The damage the tsunami did to other structures and people (without literally sinking the ground into the sea) were much more expensive than that.
I remember not being able to buy milk in Europe (Switzerland and Germany) in the 1980s because of Chernobyl. The increase in radioactivity in milk was orders of magnitude, due to fallout from Cesium 134 and 137 as well as isotopes of strontium and plutonium.
Circling back to the easy to measure part, dividing most numbers by something close to zero tends to turn out big, the ratio says fairly little about whether that's dangerous (https://www.bfs.de/EN/topics/ion/accident-management/emergen... "To date, there is no evidence that the reactor accident has caused adverse health effects due to radiation in Germany"). I'm also fairly certain neither strontium nor plutonium particles (plural) made it that far.
I find text useful for repo-specific conventions, especially skills.
Pure slop.
The irrational fear-factor associated with anything nuclear is highlighted by how much attention Fukushima, the nuclear disaster, which gravely injured like a dozen people, gets relative to the earthquake and tsunami causing it, which killed 20k people and leveled a huge stretch of coast.
This is the most Opus-looking website I've ever seen.
Thousands of times of that area is lost permanently every year to climate change, extractive industries, desertification, pollution etc. (just desertification is estimated at around 2500x that number). With nuclear we always proceed with extreme to absurd levels of caution, in no small part because radiation is extremely easy to measure - unlike almost every other contaminant we put in the environment, which are harder to quantify and our risk response is generally "oh it's totes fine". See: PFAS, microplastics, endocrine-disrupting chemicals dropping sperm count and reproductive health indicators on the entire planet, PM air pollution, UPFs, blood lead levels, military chemicals etc. etc.
But if the Geiger counter ticks too fast once everyone in a 100 mile radius loses their mind. Nuclear accidents are not fine. Obviously. The impact is dramatically, comically overstated compared to literally any other environmental concern.
Building a "NN with similar capability as the brain" is not modelling its physical structure. The assumption is not made.
You're talking about a philosophical debate whether the brain is computable, the other commenters are pointing out that even conservative estimates point to a brain-like NN requiring over a quadrillion parameters.
Around 98% of Fukushima is inhabited again, unless of course you meant the NPP itself, but people were not living in a power plant to begin with.
https://i.imgur.com/t5jDXrt.png
"Why don't we unpack malware in the kernel" - "And so the search for intelligent life continues..."
Wireshark dissectors (protocol decoders) are basically all written in C, and anyone sending packets can pick a dissector.