Yah - still feel like there are plenty of examples (inc. recent ones) where this isn’t true, not specifically with ML, but someone like Jobs comes to mind.
Conversation
People love to find counter examples to this very useful and accurate rule. I don’t know why.
1
Well you phrased it (or at least I read it as) without exception and that seemed wrong so I said so heh
1
That’s how Twitter works isn’t it.
My point of course is it rarely works, it’s a big problem and everybody knows it. But people keep trying to make it work.
Like people not getting enough sleep and saying they don’t need it
2
Now we're talkin ;)
1
Actually the I think about the initial bit - deep product people mostly come to mind when I think of successful counter-examples. I'm not sure I can think of ones where "sales" worked out. I'm sure it exists. Also I am obv biased by 20 years of tech & gaming pre-baseball.
1
1
I’ve seen great product and design people — have no clue how to hire serious engineers much less build it. Even great successes like FB struggled — and spent large $$ and time with bad tech leadership early on. Twitter — even more so.
1
1
Contrast that with Stripe. Good technical founders.
Engineering is not enough. Great ideas matter more. But no one ever had too much good engineering.
1
Indeed. There are loads of examples that fit. The hiring comment is definitely one I take to heart since I realized long ago the diff between an 85 percentile engineer and a 95th (let alone 99th) one definitely isn’t linear.
1
It’s not. And even thinking about percentiles is kind of silly.
The reasons not to hire the best though exist
* you don’t need them
* you can’t motivate them to work on your problem
* they won’t learn the non tech part that’s crucial to the job
many others. But know what’s up
2
1
I’m a fan of bad code, akshually. Facebook was built off horrible code. Sometimes that’s what’s for dinner.
They spent years and millions of dollars ripping it all out when it mattered.
Yah we agree here I suspect. MVP/vertical slice and fix it “in post”. Risky sometimes, but not always outsized.
1
Yeah it’s fine. Just know your code is 🐶💩 — it’s fine just don’t let yourself think otherwise. Bad code is fine just don’t have an ego about it
1
1
Show replies