At this point, I suspect that is just about every tech company. Your best bet is to self-host everything, no agents, no cloud services, completely locked up home network and a loaded shotgun if anything starts making unexpected noises.
I have a good friend that works for one of the giants(I can't say which one for obvious reasons but S&P 500). He's been working there for quite a while now, so far he hasn't seen what the project he works on looks like, has the repo cloned and knows what language is used but nothing beyond that. Everything is slopped together. His project is the authentication and authorization system for all the company products. In his own words "I hit Tab all day long and write 'this is intended' in the reviews, which are all ai, there is no human in the loop. This is what we are told to do by the CEO and CTO unironically. If something breaks, no one knows how any of this works since no one has seen the actual code. Our performance reviews are based on how many tokens we've used, not what we have done". I suspect this is the case in many companies now so it's not unreasonable to think that there are no actual code reviews.
I hate to be the "I told you" guy but... I told you and have been for years. And every time I do, a flock of sloppers come to say "but have you tried the claude sloppus, it's so good man, I haven't written any code in X months". Well.. Enjoy.
Not to spoil the surprise but it will get much MUCH worse. Reason: sloppers. Anyone who's dealt with security and has looked into how all the slop agents work can understand how catastrophic it is from a security perspective. The "yes" button on "I trust the authors" is what unlocks the gates of hell.
If that were the case I'd be getting a lot less "Hey, can you check this, something isn't working, here's what the AI did: ..." in complex systems but, alas, I hear those exact words 10+ times a day if I'm lucky (up from 3-4 a year ago).
The good times are yet to come though: when all the slop held together by spaghetti, duct tape, chewing gum and hairballs starts falling apart and someone needs to fix it. If the slop bubble doesn't start popping soon, in 10 years a handful of people will end up being better paid than COBOL developers are now cause they'd be the only one that know what a compiler is. Personally I'm not too enthusiastic about the prospect so if I'm lucky, I'll be enjoying early retirement and watching all hell break loose from a camper van near the sea somewhere.
I'd be more than willing to subscribe and support the project BUT, I need to address the elephant in the room: The reason why I'm against Kagi is the fact that they use Yandex(be it only for images according to their own words) and I'm sure as hell refusing to give a single cent to them. So I guess my question is: sources?
"The costs are negligible and justified when compared to all the benefits. If you look at the performance gains, the overall cost has in fact been reduced."
- Altman, a.k.a. Dory from Finding Nemo and/or Dario, a.k.a. Carl from Jimmy Neutron.
Brilliant article. This is something I've been thinking about for a while. Up until around 2020 I used to work at a company that lived off of games and the economy was, you guessed it, micro-payments. I was the one in charge for developing the system that allowed the people in charge of monetization to configure the games based on your skill to squeeze the most out of you. Suffice to say, it worked great. Fundamentally the business model for all games was identical: cash for virtual currency. Here's the catch: you never knew if spending 50 bucks would make a big difference and you had no way to measure it. In a nutshell, it almost made a difference but just not enough so your brain would go "well what the hell, here's another 50" (classic sunk cost fallacy). And the business knew that and actively exploited it. All the AI slop that is happening now is the evolution of the same thing: exchange cash for virtual currency(tokens) in exchange for immeasurable results and the inevitable "just a few more tokens". Congratulations, you've been played.
searxng works well but it's not perfect. I tried qwant for a bit but results were mediocre at best. Kagi, I've shared my views on why I'd never use it. I wouldn't use it even if they paid me to use it in fact. For the time being I'm swapping between searxng and ecosia, the latter has been giving me the best results for just about anything(and I hope their indexer partnership with qwant pays off). As for google - even before the change, results were absolute crap, not a recent thing but since around 2020, the quality of the result hit rock bottom and they have been hard at work, drilling into the rock ever since.
I assume they are called hooks, .vscode/settings.json - you can put some linters/tests which run automatically (from my understanding, something similar to git hooks, hence the reason I called them hooks). I generally hate the concept and I generally dislike vscode so... yeah.
The vast majority of devs in the last few years have either been raised as sloppers or transitioned from developers to sloppers. Programming has evolved into a blackbox where fewer than ever people know how a program works, despite the numbers of "programmers" has skyrocketed. Just a few months ago a friend had a job interview(fake one obviously) where he was asked to clone a repo and do some stuff on it. The repo contained some vs code hooks which did a ton of stuff in the background installing backdoors and whatnot. In my friend's case, that was a non-issue since he was running it in a VM but I recon thousands of others have fallen victim of such attacks and have no clue still.
Love-hate relationship here. I make and fly drones as a hobby (mostly build-since around July last year, 98% of the times, when weekends come, it's always raining). While soldering can be fun, I hate soldering big, fat cables on large pads, it's just so tedious, even with a powerful soldering iron.
> you just spit at it and it'll generate you a whole app and you don't even understand a single line
So we are going to pretend this isn't happening everywhere now? And that it isn't failing on daily basis? I'm sorry but I've been saying this for years now and is my main arguments for not using slop machines: no one writes the code and no one reads the code. I can name dozens of fortune 500 companies where "tokens used" is used as a performance metric for developers, as in, more tokens = better performance, all code is written by slop machines and all reviews are made by slop machines, developers simply add "this is intended" in code reviews.
Not only that. I was probably not even a teenager or barely a teenager when registering a gmail was not as simple as clicking "sign up". You needed someone to refer you and upon registration you got 25 referrals in return. Needless to say, entirely ditched gmail forever ago and use it as a spam mail. They can have all the fun they want training slop on that.
I won't lie, I'd be pretty curious to find out I had died too. Back when I was in uni in the late 2000s and early 2010s I stumbled upon some of your talks and thought to myself "this man has got to be the most fun person to have a coffee/beer with". Scrolling through this post,I'm even more convinced about that.
I have a node running 24/7 (I also happen to be hosting one of the dozen or so things network nodes in my city) and while the idea is great, the adoption is close to 0. In a city with a 2+ million residents, I see a handful of users, as in less than 10. Same with the things network actually.
The article feels like AI hallucinated slop. Just a quick scroll through the page:
* Sviatoslav was not a local ruler - he ruled Kievan Rus' 1500km north-east and he remained a pagan until his death, even if his mother had converted to Christianity.
* Sviatoslav was born nearly 60 years after both Cyril and Methodius had died.
* In 890 Boris was no longer in power but his firs son, who coincidentally tried to reverse the Christianity conversion and was kicked off the throne a few years later.
* " Just after the invasion of Ukraine in July 2021" check the date.
Seniors are no different and that infuriates me even more. The few times I felt lazy and let an AI do a simple function for me, all hell broke loose. I'm starting to think that many people were never that senior to begin with: Writing the code accounts for very little of what development requires and is often the easy part. Understanding the problem and finding the sweet spot/optimal compromise, edge cases and how you can break it is what has always been difficult. Seeing github explode with slop and github(microsoft/openai) themselves push even harder should be a wake up call for anyone that understands what development is: not writing the code but having someone else go through it, analyze it, understand the problem you are trying to solve and why you made the decisions that you made - that pretty much always takes a lot more time than writing the actual code. And then I see someone push 20 commits in a day, each being 5000 lines, jam packed with emojis and other slop and tell me that they carefully reviewed all of them? Yeah, that's bullshit, mate.
Judging from everyone I know, it will take people a lot of time to learn and accept a lesson from decades ago(one from before some of us were even born and I'm in my mid 30's): lines of code is a shit metric. Sloppers tend to believe that something seemingly working = production ready = good execution. And the metric, of course, is lines of code or "tokens". Until then, the list on this website will keep growing exponentially.
No, it shows responsibility and critical thinking: I don't care if it's slaupecode, alanturingai++ code or whatever other slop machine anyone uses. It's the fact that no one and I mean NO ONE is capable of code reviewing that much crap and everyone that has written any code that was meant to last knows it: pouring a ton crap of code, even as a human, is the easy part if you learn the basics. The hard part is analyzing the problem, considering and addressing edge cases, exploits, memory leaks, resources, constraints and limitations - this is what a programmer is fundamentally responsible for and this is the only skill that matters. The code is just an abstraction, which was never the issue. Sloppers wash their hands with "look at how good it works" and "look at how beautiful the code is" and it's all fun and games before a production system crashes because, as I said, no slopper or an actual programmer has the ability to review 6000 line merge request so they hit "lgtm". Or better yet, outsource the problem to another slop machine. Anyone who believes that "AI tools" are in any shape or form viable for production, clearly do not understand a fundamental concept, called propagation of uncertainty[1].
"Sooo productive", sure, buddy, and when it breaks you have no idea why it broke and ultimately you don't know the codebase. Instructions.md containing:
```
* make no mistakes
* don't hallucinate
```
Is the clear-cut evidence that we are speed running towards a technological apocalypse. Ironically, I'm in favour of it: if the house has to burn down for things to get better and people to get back to using their brains, I'm all for it.
Disagree: I applaud them for doing this. Anyone that says they've reviewed the 1000 lines of slop any AI has spit out is simply lying to everyone and even potentially themselves and has never done a single extensive code review in their life. Reading 1000 lines of text is one thing, reading and analyzing the complexity implications and edge cases in code - no chance. The "I've reviewed the slop" response is the reasons why 0 days and leaks are happening more than ever: because no one really reads the code cause "I vibe-coded it". An extensive and comprehensive code review may take days, and no slopper has ever done that. I'll get a 100 line pr and going over it can easily take hours, especially when something looks wrong and I need to test it. And it's a good reason why I'd never trust the "You are absolutely correct, apologies for the oversight, here's a revised version:"