efficax

either the registry is compromised or someone at anthropic got compromised (or did it on purpose to steal crypto). the line in the cursor disclosure about the extension not being developed by anthropic is more confusing than clarifying. amateur hour for these massive corps

They really need to show some evidence for this! Just this suggestion means I won't ever use it again

Ok so it's "AI free". But why? It doesn't justify the decision but just assumes you will understand the reasoning. I have to say: I don't. For example they are pinning binutils, gcc, and gdb to the last version they know does not have LLM generated code. I assume this means they will not be able to accept any upstream patches, but instead have to take over maintenance of those packages themselves for the forseeable future, effectively meaning development on those will stop for the purposes of AlpineBSD. Seems stupid, to me.

i think you mean almost 20 years ago old man

the rumors were about a supposed Macbook Ultra that would have a screen like the ipad pros. Still to come (or not)

still rocking the m1 studio, shows no signs of stopping

right: either it’s so hard to exploit that $1,000 isn’t enough to risk “giving away” your exploit chain or actual exploitation of this bug is impractical, either way the CVEpocalypse is overblown

The author of the piece claims playing to the click is not gridification, and I agree. Even with a click track to keep you on beat, there will still be a natural push and pull of the rhythm.

this is written so terribly i could barely read it. stop using ai to write, AI CAN'T WRITE. it can write code, it can't write a blog post

use aws or digital ocean

secret encryption is always a bit of a conundrum for a running cloud system. the secret needs to be plaintext when you use it, so you have to decrypt at least on every deploy. but that means the keys have to be ready at hand, no passphrase, somewhere an automated system can get them. that makes it very hard to actually protect them from someone with full access

like the competent programmers that built php or?

github has basically become a joke at this point. everyone engineering org that i know of is at least asking themselves if there's a migration path off there if they're not in the process of building one. Every slack has a custom emoji of github on fire. Github is so joever

FDA approval of the first GLP-1 agonist was in 2005 and many diabetics have been taking these drugs for close to 20 years, so we have a pretty good picture of what long term side effects are: mostly GI symptoms, and you can have problems with muscle loss and bone weakening if you lose a ton of wieght. otherwise they are well tolerated.

i burn through a $200 claude and $200 codex plan every week. opus 5.5 is not "step change" it's just a bit better, and the benchmarks show this. Honestly with my workflow I cannot really tell the difference between the top models, they're all pretty much the same in capability, just a bit different. You can build software with all of them, and with all of them you have to validate the change thoroughly to not ship stupid bugs

Antigravity tells me specifically "Gemini 4 Argon is currently available to Ultra plan subscribers"

Only available on Ultra, so I thought maybe I'll try it for one month to see if it's good. Well, I have Google One and for some reason this means I cannot sign up for AI Ultra? This is just one of the worst product rollouts I've ever seen

but they have slowed down. the delta in capability between model releases now is smaller than ever before and i expect it to continue to diminish until we reach some kind of asymptotic maximum, which hopefully will mean turn towards optimization and local model improvements. The current frontier models running at 10x the tok/s would be incredible for my use cases

the biggest design wart of rust (well one of them) is that support for custom allocators wasn't built into the type system from the start, with allocation osmething that returns Result

'static lifetime bounds are kind of confusing in Rust, but here goes: A 'static lifetime constraint on T says: types that satisfy T must not contain any borrowed value/content whose lifetime will be shorter than the life of the program. So suppose you have a string s, and Option<&String> q. If q has the constraint 'static, and you do q = Some(&s) this should be rejected by the compiler since a reference to s gets a shorter lifetime than 'static by default (you could do q = Some(&'static s) if s is say a string literal or a leaked box but not for a normal string). In the examples it looks like you have to use a tuple to trick the borrow checker but the concept is the same: Rust allowed violation of the 'static lifetime bounds. The drop implemenation demonstrates it: drop happens in reverse declaration order, so the string s gets dropped first. Now q has a dangling reference, a pointer to dropped memory essentially, and we get a pretty classic use after free bug.

no

one can test at the system level as well, and llms make it easier

o no, they’re seeing open and close events between my laptop and my vps and database. the fbi will surely bring the hammer down on me

they should be freaking out because every time the chinese labs or non "frontier" labs release a model that is only a few months behind and much cheaper than the openai/anthropic models it shows that they don't deserve their valuations

Engineers abandoning IDEs or did they burn a ton of cash on AI?

deepseek is from the evil east, this is from the virtuous west

Porting this to rust would be a multiyear, multiengineer project normally. And in fact you would almost certainly never do it, because it would be too expensive and you'd have to take engineers off of work building new features for ages.

Reading the code does not mean you understand the code. One lesson that experience in software gave me: I never understood the code. You think it works a certain way, until you find out that it doesn't.

What LLMs make possible is for me to say: find out all the ways this thing works. Analyze the different ways we can run this software, build a fuzzer, build property tests, and run this software in every scenario possible. Log full traces. Log all the outputs. Now, analyze each scenario for bugs. You can't do that by hand.

If we are committed to it, if we put the resources towards it and dedicate the time to it (and we could do this just by saying: it will take half as long as it used to take!), software built by llms in healthcare, finance, automotive, defense, power plans, aviation, manufacturing can all be made MORE reliable and better with LLMs... without ever reading a single line of code. The LLMS are very good at logic, by the way.

Anyway all of this reads like someone who is not actually using LLMs to build software or hasn't tried them in a while. I felt the same way in 2025. I've written 100s of thousands of lines of difficult code. You, the person reading this, has probably interacted with software I've written. For a time you would've interacted with it every time you made a debit card transaction in the united states, for example. I understand code, and care about quality, and that's why I'm all in on LLMs for code.

oracle's database is still a good product, with a lot of features that don't have analogs in open source db products. You can live without them, but still, it's a nice database.