The rule in this case would be more like "you need to show a police history with proof you've never committed murder before you're allowed to enter the supermarket".
Warsaw here. I will do everything in my power to keep my kids connected to whatever social networks they want. The state will not be able to censor the internet for my kids.
Equivocating children bypassing social media age gates to child rape is disingenuous and asinine.
The point being made is that these age gates are often easily defeated and is impossible to enforce at scale.
Moreover it allows companies to collect a huge amount of bio-metric data which is difficult for users to remove once it is collected. This data has can be and has been exposed.
1. Chernobyl disaster was not related in any way with nuclear plant normal functioning. It was "experiment of how low output reactor can go" proposed and exercised by people without knowledge or experience. It's monument of why government officials should not touch anything related to science
2. Chernobyl nuclear plant was very poor copy of safe by design nuclear plant. It had critical flaw in carbon based proton catching mechanism
3. Closed zone is currently habitable (except Sarcophagus area). You'll receive much higher radiation dose simply by flying to Norway. Nature is returning there, there're plans to repopulate it
4. This was in Soviet Union and power plant was "public". Most of the power plant in EU is public if not every single one. I don't have exact information on this topic
Please do not treat entire technology, which is safe, as a threat simply because some greedy, uneducated people from the past wanted better position and paid ultimate price for it
The Chernobyl disaster wasn't caused by a single person, as you note in your comment "proposed and exercised by people without knowledge or experience."
My employer company go through JavaScript, Java, Golang and at the end landed on Rust for srver side and don't want to change anything. Everything is Rust now, regardless of traffic.
I don't know why you think it's only usable but this is your right.
In 2016 I joined C project initially created in 1999. This project had ~7k lines in gitignore and I learned this only because one of my commit created 2h of email exchange why my code does not compile. Turns out filename was forbidden by one rule.
This one file was more sophisticated than any other file in the project.
It contained "good practices", editors/IDE files of applications dead for more than 10y, personal. /tmp directories in various names, files versioning using suffixes and comments for sections v of rules starting about in half of the file.
This project learned my to keep your own shit in local global gitignore not in project.
You are the very first person I encountered who claim LLM can fix technical debt. Usually I seem comments and articles saying LLM can only produce it. And I can agree with those articles and comments
Here's the result of a prompt that specifically looks for correctness and simplifications to make "by construction" before I added "Design bar: ensure correctness by construction rather than by convention" to AGENTS.md:
I did. Just about all of the findings were re-verified to be good and then implemented by a rigorous pipeline of clankers, and the project was improved.
If it weren't for me insisting that I manually sign off on the solution of every finding, then it would have been fully automated too.
There is a lot of noise out there from anti-AI types who never seriously tried to clean up messes with AI. If you just ask for new features and never look at the code, coding agents will make a mess. If you often ask questions about how to clean up messy code, coding agents are great help for that, too. But you have to ask.
Agents (and humans) don't know what cleanups are safe or a good idea. They can get stuck in loops when different goals (conciseness, performance, etc.) clash. Their "improvements" can break things unintentionally. Targeted improvements can be good. Giant lists of "fix everything" are bad and will break things.
Tech debt comes with insufficient tests, so you won't know what you've broken until too late in many cases.
Yeah that's why step 0 is to tell the LLM to write a test suite for the thing you want to rewrite. It'll do that, but it does require adult supervision. Once you have that test suite, you can send the LLM into an inner dev loop. Like human programmers, LLMs are great when they have a fast inner loop.
The fact that LLMs exist doesn't mean you can just stop thinking. It mean the things you have to think about will be different. You have to treat them as savants with absolutely no ambition.
There’s a trap here though: the LLMs if you let them will spend all their time maintaining that test suite, they can become quite rigid in the way they design them. So it’ll write some code, break a test, which breaks a contract so then it updates the test and the contract but updating the test also broke another contract… soon enough you’re spending all your tokens fixing contracts instead of writing code.
In a sense the LLM having a fast inner loop is its blessing and curse. A blessing because it gets feedback quickly, but a curse because it becomes naval gazing and cannot see the forest for the trees.
At least this is my experience with Sol maby other models behave differently.
I’ll be the second then. I’ve found it’s very good at creating debt but equally good at resolving it. The thing is the default mode is creating debt so you have to intentionally pay it down.
Refactors that used to take me a month now take a week, it’s very handy. You can instruct them to move functions around, change interfaces, add or remove abstractions, remove redundant authorities, untangle spaghetti, rename identifiers across a codebase, and it will return very good results.
Yes like any tool, LLMs can be used poorly. But they do enable large-scale rewrites or poor-quality code because they are capable of understanding it and enumerating its behaviors in a matter of minutes or hours. It's the kind of understanding that would take a human months or years to develop.
Add to that their ability to basically set up what is essentially a perfect testing environment when asked, and you've got a feedback loop that lets you just step back while the agent cranks out a new implementation in a memory-safe language with a full test suite and bug-for-bug compatibility. I'm not kidding or exaggerating. This stuff is possible now, people just need to look past their anxieties about being replaced.
Most short-term learning/adaptation is already handled in-context. Modern context windows can hold several books worth of text - plenty for most tasks. Everybody is already using it to adapt models to their projects through skills/instructions/guides etc. ps. I often say that after glossary-skill next must have one is update-skill-skill that threats all .md files as live documents.
Persistent weight adaptation also happens just not in real time - sessions are captured, analyzed, transformed into training data, fed into SFT/RL environments and later contribute to model updates. Takes a bit of time for the whole loop but you can't say it's not present.
There's nothing fundamentally preventing real-time weight updates, ie. LoRA-style online adaptation would be one obvious approach. It's just generally not worth doing at scale. Updating a shared model centrally gives much better data efficiency, batching, evaluation, control etc. than continuously training a separate set of weights for every user/session.
There is some work happening on narrowing that gap, for example Mistral has been pushing efficient LoRA-based customization, continuous pretraining, model adaptation etc.
I also did play a bit with activation steering – it's super cool where you extract profile for some concepts (emotional in my case) and you have effectively toggles to control "brightness/contrast" those areas (enhancing or suppressing those activation regions from profile) injecting to the model those concepts (emotions in my case) – you can do it in real time and it's fun thing to play with.
My Claude admits mistakes and then fixes them on its own all the time.
Often even without my input: "(thinking..) Oh I discovered that I misjudged XYX, let me fix that.. (thinking) (executing scripts) Okay I corrected my mistake, I had accidentily ABC."
But the current state of AI is a very good indicator were it is going. Even if we don't achieve AGI, do you think AI in its current form will not be better in 5 or 10 years and we will have this solved a lot more/better than today?
Even with AI, you still need to design the software, test it, try it out, deploy it and scale it. An proper agentic platform still needs time to be build.
And the models need to be a little bit better and for that you want human feedback/reinforcement learning
>Even if we don't achieve AGI, do you think AI in its current form will not be better in 5 or 10 years and we will have this solved a lot more/better than today?
In my sector? No, not really. Because the core problem isn't tech, but labor rights. Those will only worsen.
Yes, I don't think this will be a lot better. Newest Claude is barely better, same with the rest of the model. In the same time cost of training new models and running them is increasing so taking this in the account I don't think this'll even last 5y
Really? From what I see everyone are raising prices and even if tokens price is dropping price of running LLM is raising. LLM is highly subsidized and will never be cheaper than human
In Germany, in this case selling shovels if you want to take the example literally is selling construction machines, or renting them, versus having guest workers digging the holes.
Are you aware AI companies are subsidizing usage and have negative revenue?
And now your comparison is inaccurate. I said whole application. So the correct one would be selling/renting construction vehicles vs building entire building and here selling/renting is maybe 10% of developer income.
Digging a hole is like making small code change and if you wish to continue using this analogy then tokens are digging a hole and selling entire applications (thing Software house do) is construction company building entire building
Second obviously deliver more cash and here AI premise starts to crack. AI can't deliver anything good, costs much more than developing team (Uber proved it) and even if you use it as hard as possible and it does what it suppose to deliver max 5% increase across the board
I'm 40yo. Do you know how long I'm waiting? Only taking for an account this newest hype I heard like 4x next year and the current progress is slowing significantly over last year. Even foundings are starting to back off
reply