My point was just that neither company's engineers has the necessary knowledge to make any meaningful assessment of the other company's architecture and codebases. Both sides are the result of years of iterations to address very unique problems.

Bullshit. That's like saying one company's financial team cant assess another's finances and structure. Happens all the time.

Agreed. If someone has a deep enough understanding of engineering fundamentals, moving bits is moving bits. It doesn't matter what use case is sitting on top of it.

Also - Let's not act like Twitter is absurdly complex. Handling that amount of traffic was hard in 2008, yes. It's not hard in 2022. The app itself is what? CRUD with stream processing? Some complexity with scaling that, but it's not self-driving... that's for damn sure. Even if they haven't reached L5 yet, L4 AV is an infinitely harder problem than storing arbitrary text/images and serving ads.

> L4 AV is an infinitely harder problem than storing arbitrary text/images and serving ads.

It is but it's a different problem. Solving Fermat's last theorem is an infinitely harder problem than, say, speedrunning Minecraft in under 13 minutes but you won't see Sir Andrew Wiles doing the latter - because they are different skillsets.

edit: No idea why I typed Simon Wiles, I know full well his name is Sir Andrew!

> Agreed. If someone has a deep enough understanding of engineering fundamentals, moving bits is moving bits. It doesn't matter what use case is sitting on top of it.

VERY few developers are at this level, though. Most (>99%) are stuck in some paradigm associated with their typical tasks.

You don't just get a CS degree and just magically understand everything there is to know about software engineering. There are specializations just like in the medical field and many others.

Tesla engineers working on AI have their own domain specific knowledge they likely spent many years to get a PhD for. The same goes for Twitter engineers working on natural language understanding AI to detect certain types of posts. By your logic there is just one broad category of AI researcher and they're all the same and have the same expertise. Scratch that.. there's just one broad category of computer engineer and they're all the same and have the same expertise. Twitter or Tesla engineers could easily do a code review of Space X's code base right?

You might not know how to make a sophisticated ai for censorship, but looking at the code and infrastructure, you are likely to figure out where is the blackbox/political censorship is. And since it seems that Musk is going to get rid of most of censorship(or rebuild it in his favor) that's most likely what he's asked trusted engineers to look at.

Are you serious? do you know how software development is done? do you know what a repository is? Do you know how to look through repository to find code changes? Any computer scientist can do that there’s nothing special about Twitter I don’t know why you think that. Twitter could be created by an undergraduate CS major.

There is a reason the best practice is not to do huge PRs because it's very difficult to review. Now imagine one person has to review all the code generated by a company such as Twitter with 3k+ Engineers across multiple years. Is that even possible given the cognitive load? If so how long would it take to produce anything meaningful?

It sounds like you have at least an undergraduate CS degree. Go ahead and recreate Twitter and share your GitHub repo for us to review. No need to rush. I'm happy to wait several decades for you to complete the project.

I mean how hard could really be since nobody really has access to the fire hose and only gets showing a little bit at a time which is really horizontally scalable. I mean let’s do the math when NVME Drive can write at 4gbps so scale that to match the influx of data rate with room to spare. The point is Tesla engineers are going to look for nefarious action, not to worry about corner cases. what a non sequitur argument that is

And the code is not the hard part it’s the scaling in the architecture to handle the rate at which the data will flow.

> Twitter could be created by an undergraduate CS major.

All the codebases will look like they were created by an undergrad CS major and then have accumulated 16 years of edge conditions. And if you pull it all down and try to rewrite it because it is all overly complicated and everyone who built it was wrong: https://fs.blog/chestertons-fence/

Oh, please there’s not that many edge conditions to Twitter for God sakes. It’s not Google.

There's actually a meaningful number of Tesla engineers who previously worked at Twitter, and vice-versa. Which right away invalidates your point.

He has to start somewhere.