[karpathy]

Elon also understands deep neural nets a lot more than I think people imagine. He starts with good intuitions and mental models, but also actively asks for technical deep dives, and has very good retention. E.g. I recall teaching him about our use of focal loss in contrast to binary cross-entropy for the object detection neural net (I said it had given us a 5% bump and he asked to know more) and he understood how it works about as quickly as you'd expect a PhD student to. The fact that he can do this across many technical disciplines is impressive and borderline superhuman. I don't think people understand or would believe how low-level and technical typical meetings with him are. Just saying because I get triggered reading way off innacurate takes on this topic (original comment).

[jiggawatts]

I think what upsets a lot of the Silicon Valley types here on HN is that people just like them are being called on their bullshit and fired en-masse for it. That has got to be uncomfortable.

You know the old saying: "It Is Difficult to Get a Man to Understand Something When His Salary Depends Upon His Not Understanding It."

That's what's happening here.

That engineer that was humiliated publicly for defending a slow-as-molasses JavaScript-heavy microservices Rube Goldberg machine? Hacker News is filled with people just like him that have built near identical software in other orgs.

Understandably they're upset and are looking for any excuse to dismiss Elon's criticism of not just Twitter, but their entire industry.

[phphphphp]

I doubt you'll find many people arguing that the value of employee output is evenly distributed across all people in a company: regardless of discipline, there's always a minority who are delivering the most value. There's always the most valuable software engineers, the most valuable sales people, the most valuable executives. The problem with Elon's specific brand of this take is that it's ignorant of the real-world human aspect: Elon could pick the best software engineer he has ever worked with, and helicopter them into a dysfunctional environment, and they would struggle to deliver value.

If Elon had joined Twitter, and spent time understanding the business and environment and then excised the people he felt weren't contributing towards his vision, that would be one thing... but he has made arbitrary judgements based on absurd metrics like lines of code or willingness to show up at 1am to draw on a whiteboard, he has not made judgements based on the quality of the work or the value people have delivered.

Likewise, to suggest that a software engineer is bad because they were a part of a team that built a "...slow-as-molasses JavaScript-heavy microservices Rube Goldberg machine..." is absurd: what if that person was the only reason that it wasn't 10x slower? What if, they were the lynchpin in that team ensuring that brought everyone else up to a much higher standard which ensured that what they built was usable (even if it was bad)? You cannot judge the contribution of an individual without considering the wider context.

I have no problem with a company cutting most of their software engineers (I encourage clients to minimise their exposure to software engineers, I encourage careful hiring over volume) but what Elon is doing is... not that.

[jiggawatts]

> ...spent time understanding the business and environment...

You're probably right, but there's a decent chance that Elon's heavy-handed approach was necessary.

I've seen the "gentle" approach fail.

For example, at $dayjob a bunch of on-prem stuff is being slowly modernised into the cloud. Very slowly. Slow enough to give the dinosaurs time to play politics and protect their turf.

For example, the networks teams that are used to legacy in-line firewalls will cozy up to some non-technical senior manager with a budget and get them to approve a project to roll out this legacy technology in the cloud. That way they don't have to retrain or -- worse -- risk being made redundant.

If instead some team comes in and simply bulk-migrates workloads from on-prem to the cloud... breaking a handful in the process and just fixing forward, then it appears to be messy and crazy, but the effect is that the legacy data centre teams are made redundant virtually overnight. Now they've got no clout, no time, and no pull. They're simply walked, and will find jobs elsewhere.

I've seen both approaches, and the latter style worked better long-term.

[sgu999]

> That engineer that was humiliated publicly for defending a slow-as-molasses JavaScript-heavy microservices Rube Goldberg machine?

Citation needed ;D

[MockObject]

https://www.theregister.com/2022/11/14/musk_twitter_rpc_spat...

[roguas]

Every org has slow-as-molasses, badly designed, illogical components. It is about ratio.

Time will tell. I don't think Musk/twitter's case will set any precedent. He is too much of a character to provide broad meaningful insights into industry. Also he has accumulated a list of failures which are rarely mentioned.

[saagarjha]

If that is the case, why has he been making mistakes that seem fairly elementary on Twitter? Like, I understand not understanding a problem space and wanting to learn more. But you say he has good intuitions and mental models–I would've expected at least some basic background research before posting online. Why aren't we seeing that?

[lupire]

Has he ever make a public statement, at least a paragraph in length, explaining something technical?

It's hard to blame people based on his decades of public behavior and lying about his education, falsely claiming to have a physics degree and to have been admitted to grad school.

[3eto]

Thank you for taking the time to write and share your unique and relevant insight.

[lawrenceyan]

This is why I still regularly read HN. I appreciate your commentary.