It’s juvenile to consider all LLM assisted coding as vibecoding. I’m not going to expand here because this topic is about as much fun to discuss as politics, but coding assistant tools are just tools.
If you give a regular person a race car, they will crash it about as fast as their vibecoded app crashes. Give the same race car to a pro age it’s a different story.
To be fair, we don't really have the capacity to run satellite surveillance on each and every target we select to engage in a sneak attack.
I think sometimes people watch hollywood movies and get the impression that it represents a kind of cataloging of our military capabilities. A demonstration of what we can do to our enemies. With the underlying subtext being "don't mess with us."
I just want to gently suggest that not everything we see in movies is factual with respect to military or intelligence capabilities.
I'm an old timer. I got off the bus at Quantico in 1991. But even though I'm not in right now, I'd feel confident in betting that we don't have the capacity to surveil that many targets via satellite for, say, 1 week, prior to our attack.
(Of course, when I got off the bus at Quantico in '91 I also would have been just as confident in betting that the US would never engage in a first strike. So what do I know?)
That is true for an active war but I don't believe it is true if you have literally months and months to plan an attack. Unless of course there was no plan until just a few days before and you stupidly threw a ton of your advantage right into the trash.
I live near a military base, and there is a daycare, school, rec center, pub, ice rink, church, and grocery store, open to the public, and not managed by the military. All of it is on land owned by the military, but outside the wire.
The fact that these facilities exist on military land near a base (which a hostile government would surely argue IS the base) does not mean that the people in those buildings have it coming.
Technically the statutes of Rome forbid using human shields.
A nation state bombing US mainland bases sounds rather implausible, although I certainly would prefer that civilian infrastructure to have a minimum distance to military targets, even in the US, even if only to set the right example to the rest of the world.
I do believe there would be value in modernizing the statutes of Rome regarding human shields, which would force nation states to compile machine readable lists of school locations, so that non-existent reported childrens schools and secret childrens schools would be automatically screened.
Keeping the school secret, or reporting a school location too close to a military base would then activate the right of the international community to attack that nation, in order to prevent nation states from using elementary schools etc. as human shields.
IRGC wants nuclear ICBM's. Iran invests heavily in STEM education and physics. The whole population is aware of such goals, the whole population is aware of the adversarial relationship with the Western hemisphere. Imagine your child being allocated the school that was bombed in Iran, but before it was bombed: wouldn't you protest and ask for your child to be allocated to a different school? They risk being the first casualties when the inevitable escalation to war occurs. Clearly in this fun society of Iran, those parents didn't get a choice, and could only pray their kids get through elementary before such a foreign attack occurs.
IMHO, the most damning aspect is that proper, modernized international law clarifying the permitted action-reaction patterns around human shields could have prevented these deaths, by disincentivizing such nations from using kids as human shields.
No one is using human shields. There are just non viable targets next to viable ones. Blowing up a school intentionally because you have bad intel or incompetent staff is tan unmitigated fuckup and war crime. We also don’t give two shits about “Roman statutes”. There is no moral obligation to attack a country that has schools near military facilities based on a dead empire from a couple millennia ago.
On closer reading, this is an insane take on a bunch of levels. Your username being a well known nazi isn’t a mistake is it?
> We also don’t give two shits about “Roman statutes”.
?
It doesn't improve credibility if you openly express disdain for that section of international law that describes human shields as war crimes, which defines the concept of human shields, and then proceed to dictate that nobody uses human shields. Are you claiming this section of law is superfluous because it never happens?
BTW, my user name refers to a well known frozen pizza brand.
The well known food brand was run by a card carrying wafen-SS member name Oetker who used his influence with the party to get sweetheart deals to supply the Wehrmacht. It’s an odd choice.
I’m not expressing disdain for the part of international law that bans human shields. I’m saying that civilian infrastructure near military infrastructure is not that. The “Rome Statute” - the accepted name for the law you are calling the Roman Statutes - does not have anything in it about it being fair game to massacre civilians. The disgusting argument that these girls had it coming since they were being g used as human shields is bunk since they were specifically targeted by precision munitions intended only for the civilian target. The entire tragedy could have been avoided with the same military outcome by just not bombing the school.
I’m not going to get into the weeds on semantic details with someone who coincidentally uses a nazi username and claims a moral imperative to bomb countries for placing a school within an arbitrary radius of a defense facility.
Your takes are wholesale indefensible, regardless of any quibbles about details.
There is never a justification for intentionally bombing a grade school. Period.
> There is never a justification for intentionally bombing a grade school. Period.
let me focus on this word "intentionally", intent.
When I take the ladder I hid somewhere else under the bridge, and climb in my secret homeless bum nest, I am taking that ladder with intent. So I know intent exists. When I intend thin gs I do them intentionally, and sometimes I don't intend things like accidentally knocking over someone else's glass of drink, then I know I didn't do it intentionally.
Whenever theres a conflict before the courts, and whenever the relevant laws refer to presence or absence of intent, there will be an interest for both the plaintiff and the defendant to make claims on intent: the plaintiff might claim the defendant did such and such with intent, while the defendant has an incentive to claim such and such happened without intent. People take risks, plaintiff took risks, defendant took risks. Often both are co-responsible for a sequence of events, the law (at least on paper) is not monocausal. It is important to be able to attribute faults with causal links to damages, but an even more important role of law is to align incentives such that all parties avoid ending up in these situations. The law serves more than remediation, it serves a prevention role!
I believe it is important to prevent tragic events like girls schools being bombed.
I simply believe it is more effective to prevent harm by focusing on provable facts at hand.
How can a judge verify if something happens with or without intent?
Suppose your loved one was on the Iran Air flight that was downed, do you really care if it was with or without intent? Your loved one is now gone. Wouldn't it be wiser to leave the world in a better place, and have intent-oriented language eliminated so that the international community can promise to act strongly and swiftly when a nation state violates certain intent-agnostic conditions. Keep the (war) crime criteria objective without reference to intent, and don't do reckless stuff which can result in downing civilian air craft (whether its Iran Air or the Dutch plane above Ukraine).
Don't do reckless stuff like have a military complex, then change one of the buildings into civilian use and not marking this change on places like OpenStreetMap.
At the very least a global international list of all childrens schools, universities, etc. And unconditional permission for the international community to use ground penetrating radar, acousting sounding etc to investigate claims of colocated military compounds. International community coffers that can only be used to investigate claimed colocations (human shields). A disincentivation mechanism to prevent malicious parties constantly calling wolf to drain this coffer to prevent investigation of themselves.
Why do people believe that we somehow already have the optimum of all possible laws? Do you sincerely believe no better system of laws can be designed that prevents most of these tragedies?
To be precise:
1. I used to work in a car factory
2. Somehow a conversation with colleagues turned to my diet
3. I told them I like to eat frozen pizza
4. They started calling me Doctor Oetker half the time, and "StreetFighter" the other half of the time.
5. Thats how I got a nickname.
Whats your story, apart from shooting down ideas on how to restore objectivity in international law, because any subjectivity will result in entities believing they can get away by playing the infinite "was too! was not! was too! was not!"-game on the topic of intent...
EDIT: also I just looked up more about the history of Dr. Oetker; and you seem to conflate a couple of things
you state:
> The well known food brand was run by a card carrying wafen-SS member name Oetker who used his influence with the party to get sweetheart deals to supply the Wehrmacht. It’s an odd choice.
But Oetker didn't run this company during the war, his stepfather did. On wikipedia I read he was an SS party member though, and organized support groups for SS members and other Nazi apologetism, which is disgusting indeed...
That’s a lot of words trying to reason around destroying several hundred families by liquifying their children due to negligent intel.
Here’s how we handle this: fuck semantics. If your missile is targeted at and destroys a school, you should be held responsible. In fact, you can remove the school from the equation. If you aim a weapon at something, and fire the weapon, you are responsible for the outcome. If the outcome is a destroyed military base, congrats, you get credit for that. If the outcome is children being returned to their parents in closed caskets, you better believe you get the blame for that. Bad targeting data is the responsibility of those doing the targeting, not the people being targeted.
We already have laws around all this. The school bombing was unequivocally a war crime.
I’m not going to hear any more stupid ideas about how we can change laws to shift the responsibility for not getting obliterated by a missile onto civilians.
Does that make it not a school, somehow? Or are we cool with killing kids just because their parents might be in the military? I'm not clear what the excuse being made actually is.
It's definitely not cool to have a school adjacent to a military base.
Not saying this specific attack was justified, but whoever allowed this, let alone if it was done intentionally as a strategy, also has blood on their hands.
Where do you think the children of our armed forces go to school? There are hundreds of schools on or adjacent to military installations in the US. The only people with blood on their hands for bombing a school are the people who bombed the school. It’s really not more complicated than that.
> It's definitely not cool to have a school adjacent to a military base. Not saying this specific attack was justified
I mean, you kind of are saying it was justified, given the entirety of your focus is on justifying it. The blood is solely on the hands of the useless, dumbshit military that couldn't identify a school and avoid bombing it. And that's the charitable interpretation of their actions.
The productivity comes from not having the startup costs. You don’t need to research the best way to do X, just verify that X works via tests and documentation. I find it still takes T hours to actually implement X with an agent, but if I didn’t know how to do X it eliminates that startup cost which might make it take 3T hours instead.
The only downside is not learning about method Y or Z that work differently than X but would also be sufficient, and you don’t learn the nuances and details of the problem space for X, Y, and Z.
I find it useful to use a brainstorming skill to teach me X Y Z and help me understand the tradeoffs for each, and what it'd recommend.
I've learned about outbox pattern, eventual consistency, CAP theorem, etc. It's been fun. But if I didn't ask the LLM to help me understand it would have just went with option A without me understanding why.
In my instance, I’m talking more about using library X or library Y, not the difference between using an atomic versus a mutex. I want to learn the latter, but the former isn’t something I care about.
Ah, that's fair. I personally don't normally care about library usage as long as it's fairly well documented and effective (like Shadcn vs raw tailwind components vs chakra... I don't really care).
Yea I just use LLM agents as tools, I don’t kick whole features to them or have a cloud agent running all the time. I rarely use more than $100 in usage monthly, usually less than half that. I use tab completion a lot in Cursor and use agents to make mechanical changes or integrate features I don’t care about learning, like integrating several libraries together into my application. I also use it to write things I’ve already got examples for, like database APIs.
Software engineers who haven’t tried these tools don’t understand what they are, and vibe coders who never understood software are taking the mindshare in public because it sounds revolutionary to some and apocalyptic to others. You have to stop listening to the claw bros and try using these as tools yourself in small ways to see what it’s really about, IMO.
Agreed; as in most things, moderation is key. It is a new tool that is here whether we like it or not. May as well learn to coexist with it, but also not defer all our thinking to the tool.
> just verify that X works via tests and documentation.
No, its verify that X approach is semantically correct, architecturally makes sense, design is valid and then add tests and documentation. Basically, 80% of the work.
Most features I do are not that complex. For the complex ones I’m already feeding designs to the agent via interfaces and specific instructions.
I think people here are thinking I’m building Gmail from scratch when I’m talking about adding additional database APIs and models in the style of stuff I’ve already done twenty times in the same application. That’s easy for even the dumbest LLM and verification is not more than a code review, maybe an hour or two of labor. It’s never perfect, but as stated elsewhere I know how to program this stuff already so I can just fix it on the spot before I commit it.
I understand though, some of you either can’t use agents to code because of one reason or another, or refuse to; both are valid. I’m saying that for my job programming in a specific industry for a specific application in a specific language that AI agents actually help me out more than hinder me, and I still ship quality code in the style that matches the code base.
> You don’t need to research the best way to do X, just verify that X works via tests and documentation.
"Just verify" is glossing over a lot of difficult work, though. It doesn't just involve checking whether the program compiles and does what you wanted—that's the easy part. You should also verify that the program is secure, robust, reasonably performant, efficient, etc. Even if you think about these things, and ask the tool to do this for you, generate tests, etc., you will have the same verification problem in that case as well. The documentation could also be misleading, and so on. At each step of this process there will likely always be something you missed, which considering you're not experienced in X, Y, or Z, you have no ability to properly judge.
You can ignore all of this, of course, which majority of people do, but then don't be surprised when it fails in unexpected ways.
And verification is actually relatively simple for software. In many other fields and industries verification is very impractical and resource intensive. It doesn't take a genius to deduce the consequences of all of this. Hence, the net effect of these tools is arguably not positive.
Exactly. "Tests pass" and "code is secure" are just different things. AI code makes that gap worse.
I run static analysis on mixed human/AI codebases. The AI parts pass tests fine but they'll have stuff any SAST tool flags on first run — hardcoded creds, wildcard CORS, string-built SQL. Works in a demo, turns into a CVE in prod.
And nobody's review capacity scaled with generation speed. Most teams don't even have semgrep in CI. So you get unreviewed code just sitting in production.
The "10x" is real if you count lines shipped. Nobody counts the fix cost downstream though.
They aren't idiots, they are evil. They know what they are doing; enriching themselves and hoarding political power and resources. Claiming these folks are dumb rather than evil propagates the idea that we should give them some sort of leeway. In fact, we should have sent these clowns to prison 5 years or more ago.
I don't think they even care about enriching themselves. I don't think they care about the money. I don't even think these little projects move the needle much.
Yes. Many of these people are rich enough that they could stop working today and their descendants could maintain their lifestyle for generations. It's about power, and leveraging that power over people they don't like, for petty reasons.
Those who aren't rich enough... yes, just pure evil. Their parents didn't hug them enough or something, and now they think that "owning the libs" is a life strategy. It's pathetic. I'd feel bad for them if they weren't causing so much harm in the process.
> Training on models provides experience that could never be gained on real ships for the simple reason that neither shipowners nor local authorities would allow such risks to be taken.
I was trained onboard a real ship. It took about a year to become qualified as officer of the deck. There are also sims now which are probably better.
I don't understand your point. Are you saying that it's possible to get adequate training on a real ship, and using the fact that you managed to do it as an example?
Of course it's possible to train on a real ship. But the sailors who trained on models have certain specific experience that you will hope never to get. As for the things you experienced and they didn't, they can catch up with those during their first year on a real ship.
These small models won’t train you in maneuvering skills directly, but rather in how to think about maneuvering and the mindset you need in traffic. A sim or a real ship gives you actual maneuvering skills in addition to that experience. You could sit in a classroom with models of ships of a desk and get the same experience as this place, in my opinion as someone who did this work. Obviously you don’t want to crash a real ship, but you don’t want to be crashing sims either.
I like Cursor a lot more than Claude Code. It works better for me overall. I like the way they integrate it into the IDE so the agent is my tool rather than a 'partner' or something like that. I'm pretty sad that they lost some engineers, I hope these folks weren't integral to Cursor in any way.
Some people call themselves vegans but will still use animal products that they feel are ethical. Also, some vegans do occasionally use animal products just because they want to.
I don’t think it’s a conspiracy but it’s weird that the vegan topic even came up in this article because it is immaterial to the main topic.
The only issue to beat in mind is that visual inspection is only about 85% accurate at its limit. I was responsible for incoming inspection at a medical device factory and visual inspection was the least reliable test for components that couldn’t be inspected for anything else. We always preferred to use machines (likes big CMM) where possible.
I also use LLM assistance, and I love it because it helps my ADHD brain get stuff done, but I definitely miss stuff that I wouldn’t miss by myself. It’s usually fairly simple mistakes to fix later but I still miss them initially.
If you give a regular person a race car, they will crash it about as fast as their vibecoded app crashes. Give the same race car to a pro age it’s a different story.
reply