This $9BN valuation is highly questionable. They can make a lot of money and also lose a lot of money very quickly if the momentum evaporates.
It only takes Microsoft to destroy them with VS Code + Copilot and to further lower prices for longer (and they can afford to do that for years).
This sort of hype happened with Clubhouse ($4B valuation until the users stopped signing up.), Hopin ($9BN until the pandemic ended) and Inflection AI ($4B and no-one uses it after the hype).
There really is no lock-in case for Cursor and users can easily cancel and switch back to VS Code.
I would sell at this point, before this bubble pops and the competitors begin to gain ground.
Any such ship is going to grow its own food. They may need to carry some extra matter to compensate for accidentally or inevitably lost biomatter. But otherwise they would be able to recycle the same biomatter repeatedly to create food. The only limiting factor is energy. You'll need to add energy in every cycle. Recycling isn't going to happen for it.
Of course, all of this comes with the caveat that nothing like this may exist yet. Sealed mass ecosystems exist in glass jars, but something on the scale of a ship is going to be an entirely different ballgame.
You can tell that MCP (model context protocol) was designed by those who haven't thought about security or looked at a rigorous RFC (request for comment) document and instead done it for the 'vibes'.
MCP has the equivalent security as the over-hyped and still error-prone JWT (Json Web Token) standard.
Both are horrific 'standards' designed to cause lots of security incidents.
You are more than correct. Lots of these companies struggling for VC money end up having to rebrand and scream about being 'AI-powered' as the last chance for survival.
Somehow they are also believe they are 'AI companies' contributing to AI research all of a sudden, but are just an API call away to someone else's AI model.
Like previously when everyone was an 'internet company' then a 'technology company', then 'robotics company' now an 'AI company' and soon a 'quantum computing company', then they really are confused on what they actually do.
1. Work smarter, not harder.
2. There is *zero* loyalty as an employee. If you have no shares, leave.
3. Start investing right now.
4. Build a startup at least once.
nah, working harder is what “THE MAN” wants you to do. you should do everything else in life harder than you work. the worst advice that is given to young people is to work harder and there is nothing close 2nd
> Also, my personal experience with LLMs fixing compilation errors is: when it works, it works great. But when it doesn't, it's so clueless and lost that it's a complete waste of time to employ LLM in the first place -- you are much better off debugging the code yourself using old fashioned method.
Or just 'learning the Rust syntax' and standard library?
As you said, LLMs are unpredictable in their output and will can generate functions that don't exist and incorrect code as you use more advanced features, wasting more time than it saves if you don't know the language well enough.
I guess those coming from dynamically typed languages are having a very hard time in getting used to strongly typed languages and then struggle with the basic syntax of say, Rust or C++.
Looking at this AI hype with vibe-coding/debugging and LLMs, it just favours throwing code on the wall with a lack of understanding as to what it does after it compiles.
This is why many candidates won't ever do Leetcode with Rust in a real interview.
> I've been thinking about this over the past few days - If AI is as capable as people here claim - enough to replace junior or intermediate developers - could it be used to maintain semi-abandoned open-source projects?
I don't think AI agents can generalize reliably to maintain complex open-source software, better than the authors. Those who say that are out to sell you an AI product or something.
The benchmark I would consider is the maintenance of the Linux kernel and AI is far from approaching to the level that it can maintain correct critical code that is trustworthy enough to being merged into mainline. But we'll see.
> Overall, not a very difficult problem and I can probably do it myself given a few days. Human engineering hours are quite expensive though and for such a low value task, I would assume an AI agent would be more cost effective.
Software maintenance is indeed expensive for a reason, however we still can't assume that an AI agent can help deal with autonomous maintenance on OSS projects; large or small or even in [0] and [2].
Technical debt is also a cost and many of these abandoned OSS projects don't even have any tests in the first place, including both 'hydroxide' and the protonmail client in [2].
So how do you know you if either project really works or not? Even if you use AI, what if it introduces a new bug in its implementation? How would you know beyond 'it works for me'?
In fact, autonomous AI agents just risk introducing more technical debt in OSS projects and increases that cost rather than reducing it.
It only takes Microsoft to destroy them with VS Code + Copilot and to further lower prices for longer (and they can afford to do that for years).
This sort of hype happened with Clubhouse ($4B valuation until the users stopped signing up.), Hopin ($9BN until the pandemic ended) and Inflection AI ($4B and no-one uses it after the hype).
There really is no lock-in case for Cursor and users can easily cancel and switch back to VS Code.
I would sell at this point, before this bubble pops and the competitors begin to gain ground.
reply