This is true, and what Yahoo did in its early days was definitely impressive, but the talent that did it moved on, and was never replaced.
When I joined in 2013, I encountered a lot of systems/processes/whatever that would have been impressive in 2003, but hadn't been updated or replaced in 10 years, and the culture inside Yahoo was still stuck in a weird not-invented-here bubble where a lot of people thought their 2003 tech was still impressive.
Turns out, if you gut the engineering organization over a decade, you end up with a B team, and you're completely unable to attract the A players necessary to keep innovating.
There's a lot of criticism in this thread that Marissa didn't do anything worthwhile, but she sure as hell tried. All those acquisitions? Almost all of them were acquihires, in a desperate attempt to get A players. Most of those didn't stick around very long though, because there's only so much shit and technical debt you can stand, and then the company is out the money, and has very little to show for it. It could have worked, but it didn't.
I believe we fought many of the same battles then. I think every Yahoo engineer at that time can curse ylock. If you where good though you looked at it with the WTH ylock did and thought hell it wasn't that bad for 20th century or early 21st, unfortunately there we where in early twenty tens constained and just trying to get a 64-bit kernel.
It's worth note that Facebook also had some similar challenges. They had so much in PHP and if you are an idiot and think just rewrite well then you would be doomed to failure. The advantage that Facebook had a few years back was that the didn't go through a period of removing tech talent in the same fashion that Yahoo did. They also had cash, a lot of cash. They could afford to pay for things like HipHop. By the time Yahoo needed a tech refresh they had already burned the cash and caused the talent to leave.
I know that sounds bleak, but at the same time while I was there after MM I did find really great talent. Some of that existed before she was there. I can think of several folks on tech47 project and they know who they are. They don't really need to be told they are talent. At the same time I can think of many new guys that brought it on. I still think one of the better talks I had was a with the new guy that wrote some of the Android network stack. Without telling names, he was describing the issues between dropping WiFi and dropping cell. One is easier then the other, if I drop WiFi and still have cell I can continue. If I drop cell I can only hope for WiFi. Either case is hard because the stack needs to understand an IP address shift and adjust accordingly. When you have that talent that understands and can code it, then you get things done. Unfortunately with everything else in the valley getting a quick hit and retaining are two different things.
Facebook also has one single product, whereas Yahoo has I don't know how many products, and they're all completely different, on sort of different stacks, in different silos and budget orgs, and all slightly different such that unifying them is an enormous task.
If they had had the engineering talent and budgets to build the kind of shared infrastructure and fundamentals that for example Google did, they would have been in a very different spot.
This is true, and what Yahoo did in its early days was definitely impressive, but the talent that did it moved on, and was never replaced.
When I joined in 2013, I encountered a lot of systems/processes/whatever that would have been impressive in 2003, but hadn't been updated or replaced in 10 years, and the culture inside Yahoo was still stuck in a weird not-invented-here bubble where a lot of people thought their 2003 tech was still impressive.
Turns out, if you gut the engineering organization over a decade, you end up with a B team, and you're completely unable to attract the A players necessary to keep innovating.
There's a lot of criticism in this thread that Marissa didn't do anything worthwhile, but she sure as hell tried. All those acquisitions? Almost all of them were acquihires, in a desperate attempt to get A players. Most of those didn't stick around very long though, because there's only so much shit and technical debt you can stand, and then the company is out the money, and has very little to show for it. It could have worked, but it didn't.