criley2

I think you just accept the risk, same as not doing code review. If you're doing something Serious™ then you follow the serious rules to protect your serious business. If it doesn't actually matter that much if prod goes down or you release a critical vulnerability or leak client data or whatever, then you accept the risk. It's cheaper to push out code quickly and cleanup the mess as long as stakeholders and investors are happy.

And I think this analysis suffers from ignoring the increased environmental cost of creating a new electric vehicle. It takes 15,000 to 20,000 miles of driving for the average EV to "break even" with new ICE vehicles due to the much worse environmental impact of creation.

Reduce, reuse, repair, recycle. Vehicles should be reduced first (drive less), reused second (keep using existing one), repaired (keep using existing ones), and finally you can recycle as best as possible.

All of this should happen prior to replacement. If you replace an existing ICE with an EV, the EV not only has to catch up with a new ICE, it actually has to "catch up" with your existing already made working ICE that has no new cost for construction. That's much worse than 20k miles because the cost of building your existing car is sunk. It could take as many as 50,000 miles of driving to break even against your existing used car.

Consumerism and early replacement of working goods has always been the mortal enemy of environmentalism.

The problem with reading vs writing is building "true understanding". If you are reading code at a high level and building a perfect mental model and reading every bit of code, then you're doing it right. But many folks see finished code and get "LGTM brain" and don't necessarily fully think out every command on every line, leading to poor understanding of code. This is a huge drawback in LLM-assisted coding. Folks re-read code they "wrote" and have no memory of it at all.

In the edit experience I am using, the LLM provides a git style changelog where I can easily compare before/after with a really detailed diff. I find that much more useful than giant "blobs" of code where minor differences crop up that I don't notice.

The other massive drawback to the out-of-codebase chatbot experience (and the Edit With Copilot experience IS a chatbot, it's just integrated into the editor, changes files with diffs, and has a UI for managing file context) is context. I can effortlessly load all my open files into the LLM context with 1 click. The out-of-editor chatbox requires either a totally custom LLM with various layers to handle your codebase context, or you have to manually paste in a lot of context. It's nonsense to waste time pasting proprietary code into OpenAI (with no business agreement other than a privacy policy and a checkbox) when I can get Copilot to sign a BA with strict rules about privacy and storage, and then 1-click add my open files to my context.

Folks should give these new experiences a try. Having claude chatbot integrated into your editor with the ability to see and modify your open files in a collaborative chat experience is very nice.

Using the front end of these websites to work in your complex codebase is very challenging. I use the chatbots for higher level questions about libraries and integrations, not for specific implementation details in my codebase. But without a data agreement in place, you shouldn't (or maybe can't) paste in code, and even if you could, it's an inferior way of providing context in comparison to better tools.

However, I do use copilot+vscode with claude 3.5 and the "Edit with Copilot feature" where it takes my open files, plus any other context I want to give it, to drive changes to my files, has been surprisingly good. It's not really a time saver in that the amount of time I spend verifying and fixing or enhancing the result isn't really faster than me just writing it myself, but I still find benefits for brainstorming, quickly iterating on alternate ideas and smaller refactors, and overcoming the "get started" hesitation on a seemingly complex change. I'm to the point where it can absolutely add tests to files using my existing patterns that are well done and rarely need feedback from me. I have been surprised to the see the progress because for most of the history of LLM I didn't find it useful.

It also helps that I work in a nodejs/react/etc codebase where the models have a ton of information and examples to work with.

In all my circles, I don't know a single "X" user. Not even one. Anecdotes are funny like that.

Facebook monthly active users: >3 billion

Instagram monthly active users: >2 billion

Threads monthly active users: ~275 million

If they're cheating, they're doing a poor job of it.

Basically what Wolfram Alpha rolled out 15 years ago.

It was impressive then, too.

In my opinion, there is absolutely no way a junior / SWE1 could greenfield an app by doing anything other than following a youtube tutorial or similar step by step guide, and would have basically no understanding of the decisions made throughout. They would have no real experience with style, linting, no good opinion on libraries, no knowledge of accessibility, security, usability, nothing. They would end up with a basic 'hello world' application that would likely fall apart before you could even get your business logic in there.

If you have a SWE1 that can literally just greenfield an entire application and do it at a very high level, then they are a unicorn / 10X developer operating at a ~SDE4 level already.

I personally do not think the average SDE1 can meaningfully contribute to a quickly moving and efficient team. By the time they are even able to contribute regularly, they are SWE2/mid in my book.

I do have much less experience in giant beaurocratic slow moving organizations so I don't have an opinion on how all of that rigmarole works, I'm speaking more to highly focused small teams delivering at a high level quickly.

I don't personally agree with the idea that senior is basically the end point of writing high level code. Perhaps it's just a difference in terms but I think of seniors at the high end of their capacity as being able to release high quality features. But if I needed a whole new app greenfielded, would I trust a senior to not only get the job done, but execute at a high level, with a clear vision, clean design system, beautiful code base, all the right choices of libraries, following all the best practices, linting, style, etc? Of course not.

To me, a staff engineer can deliver that application at that level, and can lead senior engineers in maintaining and enhancing it.

We also consider a level above that, the principal engineer, who is primarily focused on higher level and longer term thinking and less about individually jumping into tickets and features. They re-envision the entire approach to implementing the business logic and scale the business in ways that seniors are mostly helpless at. They do work through others much more, and I do personally see Staff Engineer as the highest level where you primarily are still coding day to day.

Humans have an uptime of 16 hours, but generally speaking can only operate professionally for an 8 hour span. The other 8 hours are spent mostly doing personal maintenance and refueling their physical and mental capabilities.

Oh, and humans have traditionally banded together to punish those who ask for more than 8 hours a day or 5 days a week, and have even historically gotten very murdery over the subject. Buyer beware!

I'm not sure where you learned that the US legal system is against religious discrimination in private organizations, but it's not strictly true.

Many religious organizations in the US openly discriminate against people who are not their religion, from christian charities and businesses requiring staff to sign contracts that state they agree with/are members of the religion, to catholic hospitals openly discriminating against non-catholics based on their own "religious freedom to deny care". One way they can do this is an exemption to discrimination called bona fide occupational qualification suggesting that only certain people can do the job.

In a more broad sense, any private organization with limited membership (signing up vs allowing everyone) can discriminate. For example some country clubs discriminate based on race to this day. One reason for this is that the constitution guarantees "Freedom of Association" which includes the ability to have selective membership.

Really goes to show how bad top-tier LLM's are at rather basic tasks like creating a clue out of popular media. This should be among the core competencies of major models with tons of available training data and simple summarization requests.

> want free software to run website

> want free plugins to add functionality to site

> absolutely will not hire developers or pay for software or plugins

> how dare the free plugin for my free software not be coded to the highest standards

We're in the "add billions in dollars of datacenters" and "spin up nuclear power plants" level of infrastructure needed for their growth, so if anything, I might be underestimating the costs if they grow as much as Sam claims they can.

Didn't Sam ask TSMC to spend $7 trillion on new fabs? By comparison, $18 billion/yr spend seems very small.

$11 billion revenue in 2025, but how much in costs?

In 2024, they are projected to bring in $3.4 billion in revenue and lose $5 billion dollars.

They just had a massive fundraising round for $6.6 billion which at the current costs and growth is what, 6-8 months of spend?

If they bring in $11 billion in 2025, I expect them to lose at least $18 billion dollars. Good luck!

This doesn't make sense on many levels. "Hospital IT" does not code the hospital EHR systems, just like the airport doesn't code flight management systems.

These are life-long software engineers, just like others reading this comment, using the best tools at their disposal to engineer lifesaving software. They're not using "regex" to develop algorithms for monitoring patients (???), and frankly that suggestion is so wild that one has to assume you don't know anything about algorithm design at all.

An LLM literally hallucinates incorrect answers by design and struggles to get extremely basic math and spelling correct.

You're welcome to put your literal life in the hands of a hallucinating english generator, but when it comes to healthcare, I want a "0% LLM" policy. LLM's will be the cheap things that offer substandard care to poor people, while the wealthy and elite enjoy personalized and human-centered care.

In Linus's follow up video, he admits it is true and even agrees with Youtube's takedown. He states flat out that he knows other youtubers have been banned for teaching folks how to get around youtube's business model and it was a risk that wasn't worth taking for his company.

Just like how I made the analogy with floatplane, he made a very similar connection.

https://www.youtube.com/watch?v=apdZ7xmytiQ

"Youtube deletes LTT video that explicitly teaches users how to block ads on Youtube" might be a more honest title, but less clickbait.

I mean, that's their business model. Watch ads or pay for premium. Do you really expect them to host videos on how to cheat their business model?

If I went on LTT forums and wrote about how to steal floatplane (LTT's subscription service) content, would they leave my post up?

They say that copilot is great but the chatbot was bad because it lacked context, missed nuance, and frequently gave correct-sounding answers that actually weren't great.

That's exactly my experience with copilot. For simple text transformation (see this pattern? Repeat it. See this comment? Make another. See how I'm doing this? Finish it) it's ok but when asked to write any amount of code beyond that, it only writes mediocre good-seeming code that when truly analyzed within the context of the codebase, the feature, the libraries in use, the framework and language... suffers from all the same problems. The number of times I've had to walk a junior through some bad react code that an AI wrote and ask them "what do you think this is doing" and have them shrug at me is ridiculous.

On the bright side, those of us at businesses that still care about quality have a lot of work on our plates cleaning up the AI slop getting pushed into code reviews, so yay job security

Many phones had great waterproof ratings with a headphone socket. Nearly all phones today are thick enough for sockets, a 3.5mm would fit fine on my Pixel 7 pro. The thinness wars ended when too many people sat on their phones and bent them.

>You say overly-regulated, I say sufficiently regulated. Maybe, just maybe, those companies should not launch here

Sure! Sounds like we're in total agreement that the comment I replied to suggesting that the EU was "too big to ignore" is false, and that their regulations do in fact make it so that some products stay out.

>Unlikely, it's a huge and valuable market

They said the same thing about Facebook News and Canada.

Many tech companies already refuse to launch in overly-regulated markets, or launch years late with "government special" versions of the software that meet the extreme demands of the EU or China or whoever.

Here's the thing, it's been shifting for all 10,000 years of civilization.

I have long since believed that maintaining a token "acceptable" social media presence is much more powerful than abstaining with regards to privacy/anonymity. The two suspicious things are overuse/misuse of social media, and no social media at all.

It's like having a honeypot of easily found money in your home (barely hidden in a sock drawer, etc) so that a potential thief will be satisfied with the find and not dig deeper.

Stories like this only solidify the power of maintaining a controlled social media presence.

We call that anatomically modern human or early modern human, and we have fossils going back over 300,000 years.

An infant from 300k years ago, if brought to modern times, should grow up and be capable of everything modern humans are.

Any of them? Coal is far more dangerous to human health and life than nuclear.

Pick your quasi-legitimate source, but they all agree that coal is ~10X more deadly than nuclear.

> Coal-fired energy chains are estimated to cause 12 times more deaths per gigawatt-ampere-year than nuclear energy chains, and coal is estimated to cause 820% more deaths per terawatt-hour of electricity produced than nuclear

The way we use algorithms to pump offensive fake news content to everyone is shocking. It's no secret that Meta (facebook/instagram) pushes sexual content to minors and I think Meta may be the organization pushing the most sexual content to children in human history. And yet, we just ignore it?

The news sources that get pumped into Windows are similarly as bad. Celebrity sexual content, openly fake political propaganda, scare mongering nonsense, just truly depraved garbage.

I don't understand how these organizations reputations survive this. They're not just polluting our own systems with their depraved content, they're intentionally* targeting children with wildly inappropriate sexual content.

This might seem radical but I think there should be severe criminal penalties for corporate executives who allow this kind of thing to happen. Until billionaires are truly brought to justice for the media hellscape they design to pad their billions, we will be completely at their mercy. I harbor few good feelings towards China and their system but they know how to disappear a billionaire who is "too big to be good" and magically the same services' algorithms don't push sexual content to children there. Tiktok* has a completely different algorithm and pushes completely different content. Maybe they figured it out.

I'll take my 5+ weeks off per year (with strict 40 hour maximum weeks and strong work life balance) and you can have your self-employed 2 days per 30 years = 0.06 days off per year. That's 32 minutes of personal time off per year.

>The article paints those measures as if they were a good thing and now things have regressed.

I got a totally different vibe: Japan is notoriously xenophobic and maintained extremely strict regulations on visitors that were far more stringent than western countries.

So the framing here sounds like a Japanese perspective: "we didn't want you to come, and we started to let you back in, and look what happened to us" which is probably a popular sentiment there.

Imagine you come from a small town where there are no parking lines at all, and everyone efficiently parks on unmarked blacktop in a respectful way.

Now imagine you go to a big city where they have a bunch of lines in the parking lot and people only half use them correctly, parking over the lines, diagonal, etc.

The existence of lines doesn't guarantee good behavior. The absence of lines doesn't guarantee bad behavior.

This is the argument I see for javascript-only folks who don't necessary enjoy using "the worlds most bloated javascript linter"

For the record, I am a Typescript enjoyer and I use it in my personal projects as well as professionally, but even I can admit that it's not automatically superior to javascript and it has a number of really frustrating and time-consuming downsides.

It's very easy to type the args and returns of a function and protect callers, but it's much more challenging to work with types between libraries and APIs all together. Lots of `as unknown as Type` or even the dreaded `any` to try and cobble the stack together.

I mean, we absolutely did "Free PC Checkups" and obviously we knew how to tinker with startup and do virus and adware removal. These aren't some magical skills that only the rarest of employees knew, it was absolutely basic nerd knowledge that everyone who even applied knew front and back.

Back then of course adware/malware scanners were huge. We certainly did hook computers up and start a bevy of automated scanners and tools and fix performance issues. And every single time it cost a lot more labor and a lot more money to the customer. But it's up to them and a lot of people didn't want a reformat or had a more expensive computer that wasn't as economical to replace.

Believe me when I tell you that productivity wise, a guy reformatting 5 computers at once in an hour flat is lapping anyone who is spending multiple hours per computer running those scanners and tinkering setting by setting.