Comment by opnitro
5 years ago
I don't think this is true. As the author notes, he doesn't have any ability to scale due to simplistic design decisions he made. As the author notes, the hard part of this business is not "write a webserver that takes a program from a user and runs it"
Indeed, a great deal of current security work is on the problem of not running other people's code.
Yeah, the entire premise of this kind of company is asking strangers to give you arbitrary code and then running it. I imagine there ares some important design decisions there that are not trivial to replicate. At it seems the author made _no_ attempt to replicate them, as he said anyone could knock his server over easily w/ a fork bomb.
I thought these type of interpreters ran in the user's browser. Cross-compile the interpreter to JS or webasm, stream it to the user after they click on which language to use. Built-in libraries could be streamed on-demand the same way or they can be bundled with the interpreter. It would solve the scalability and security problem.
Even so, the fact that Repl.it felt so threatened by it as to threaten legal action and bully someone into taking it down speaks volumes to its viability as a competitor. The inability to scale can be fixed - probably not trivially, obviously, but it's very much a possibility.
Or it speaks to Repl.it's CEO's lack of understanding of the technical differences. I wouldn't assume anything about the viability based on this reaction.
They should feel threatened. There's nothing particularly special or novel about what Repl.it is doing, also very little in the way of specialised knowledge required to build a competitor (note that I'm not trying to belittle the individual involved here). That they got $20 million to pursue this product in some ways surprises me given the relatively low barrier to entry.
If the code is well-documented and everything is nicely set up, you just need the right person who has access to VC and an untapped market (e.g. China) to pick it up and spin off from there
Or at least it would save up a lot of boostrapping cost. Otherwise this whole thing indeed makes no sense.
> If the code is well-documented and everything is nicely set up, you just need the right person who has access to VC and an untapped market (e.g. China) to pick it up and spin off from there
I think the idea is that it's trivial to take what you've described and add "a small amount of work that an early-career engineer (even if talented) describes as easy".