It's not dumb luck. MS made a fortune from marketing and business politics, not from cool technology.
MS was never primarily a tech company, it was a sales-and-marketing-and-M&A company that happened to sell tech products.
So the 'Does this company have a future based on this product?' question is misleading.
The important question is - how are they marketing the product? What relationships do they have? What relationships are they building? What's the sales and marketing culture like?
MS are poster children for this. Time and again they've marketed the crap out of technically mediocre products with - mostly - great success.
AirBnb weren't exactly light on the marketing either. (Etc.)
If you completely dismiss their product then you would have missed Microsoft because it's highly likely that at that point their marketing etc wasn't great. That's the trap the article is warning against.
The more expert you are at something the more likely you will dismiss things. PG recognises this bias and is trying to keep his mind open. You have fallen straight into your own trap which is narrow mindedness presenting itself to you as expertise.
They did have a product and they did go on to build many products. Many of them are far from crap and solved really difficult problems including running on all kinds of hardware, maintaining backward compatibility, rewriting their main OS from scratch. Microsoft is not Coca Cola. PG is trying to see that in other companies at a very early stage. (not giving him a free pass - I think he still has other blind spots that annoy me - I have many more no doubt)
edit: No doubt, it's probably that combination of being an expert at something but having a somewhat open mind that got PG started on his adventures in the first place.
I have no idea if I would have dismissed the product. In retrospect, MS BASIC had a 4K footprint which also included floating point. That certainly made it stand out from the competition.
But I wonder if an investor might have noted that Gates had an interest in aggressive market-building and income generation, as suggested in the scrappy and public fight with the tape sharers and the eventual battle between MITS and MS over royalties.
I wasn't paying attention at the time, but looking for similar patterns today - in, say, Uber - suggest that there's a clear interest in aggressive and cut-throat behaviour, which probably bodes well for future investment returns. (If that's your main interest in a tech company - it's not mine).
I certainly don't think MS were crap at everything. But there's also no denying that the user experience with many MS products was - and still is - shockingly bad.
Office and XP were good-enough, with a few nice tweaks. But you have to balance that against a long string of horrors (insecure VB macros, IE6, the ribbon, Win 8 tiles, Win 95 series bugs, and on and on.)
So historically, the focus has always been on sales and using legal and corporate lock-in to sell poor-to-good software.
It clearly hasn't been on creating state-of-the-art super-products and assuming that quality will sell itself.
So I think the point stands - look at business culture and management attitude to market building as a primary predictor. Product quality is not a predictor, as long as it's good enough.
Being slightly better than average doesn't hurt, but there's no commercial need to be a technical marvel.
> It's not dumb luck. MS made a fortune from marketing and business politics, not from cool technology.
MS was never primarily a tech company, it was a sales-and-marketing-and-M&A company that happened to sell tech products.
Could you draw overlapping things without repainting using 1985 technology?
That is, an IBM PC running at 4.77 MHz; with 128 KB of ram (or 640 KB if you're lucky). The OS would have been dos at something like version 2.x
Some people[1] used Wordstar as their text editor.
You had limited multitasking -- TSRs maybe.
I'm not sure what programming language they used but possibly Microsoft C and assembly.
To add a little more color to this, you had to be able to quickly repaint just a region of a window. And you don't have a bitmap of the window's current contents because that takes up RAM and you'd need CPU time to update a copy every time the window updates (slowing down any animations).
I don't know how it's handled these days, but when I was first playing around with Win32, you'd receive events telling you a region of your window was invalidated. You were expected to figure out what controls and images belonged in that region and just repaint those items because of the CPU load.
Sure, because that was the obvious standard technique in 1996. But for it to become the standard, someone first had to invent it. (And Swing in 1996 probably ran slower than Mac OS in 1986, despite Moore's Law.)
