Comment by mchinen | Hacker News Reader

Comment by mchinen | Hacker News Reader

Parent Full thread

It's interesting to hear your experience at this point in time. Mine was later but noticeably different.

I started coding in the 90's as a teen releasing shareware, with my first gig at Adobe in 2001, where they paid me $14/hr as an intern. Even though this was rough times for tech, me and my CS peers at UW felt a lot of optimism and enthusiasm about what could be built because the framework still supported idealists and we cared less about the economy (and we were definitely naive). Both the researchers and entrepreneurial types were very curious about inventing new paradigms.

When I talk to younger people now in CS, the enthusiasm seems to be split for 'pure researcher' types and 'entrepreneurial/builder types', with the latter having interest concentrated on what is booming (like AI), and more about what can be built but what will be able to raise large sums or attract more users. I'll caveat this with I don't know to what extent the people I talk to do have a bias, but I do wonder if there are less people willing to explore new frontiers now.

One major difference between now and then is that the fraction of US market cap that is tech, and to a lesser extent, tech's importance in the economy. I wonder if this established leader position somehow could make people less optimistic and willing to explore?

> When I talk to younger people now in CS, the enthusiasm seems to be split for 'pure researcher' types and 'entrepreneurial/builder types'

In my view the pure research types who land jobs have always been maybe 1%? UW is research heavy, don’t forget. And at the time I was at Microsoft almost no one had plans to quit and start their own company. Before I started there in 1996 I figured about 80% would—-an early indicator of how distorted my view was from press reports and my own biases. It seems to me that these days the vast majority are happy to be drones. Or am I way off?

In my time most of the employees were very smart students who just kind of fell into programming and were hooked in by Microsoft’s pioneering efforts at sending recruiters to colleges (rare at the time) and especially less-obvious choices such as Waterloo and Harvard.

What kind of shareware did you publish?

It makes sense. I should say I'd estimate 70-90% of the class didn't want to work their own ideas, but would be happy to work at a big company forever, and this was the mindset going into the CS program. A fair number of people within my circle of folks that hung out at the CS lab and lounge and discussed ideas eventually launched successful companies, and I'm probably overindexing on these types with survivorship plus hindsight bias, since they were the most memorable to talk to.

I interned at Microsoft in 2005, and it felt like most people there were just looking for something stable, with a side of Office Space-esque exit thoughts. The culture and scale was different in ways that seem weird to me today - One example is that Bill Gates invited all of the interns to have dinner at his house, which is something that is probably reserved only for startups nowadays, if even.

Re: shareware, I released a crappy checkers twist game for Mac, which was also very educational, and I was proud to have it on the MacAddict Magazine CD. [1] https://www.macintoshrepository.org/5196-checkerwarz

By contrast, in 2009 the several hundred interns were bussed in a police escorted motorcade along the closed down highway from Redmond to the Pacific Science Center for a special showing of some Harry Potter film and catered dinner, after which complimentary Xbox 360s were distributed. It felt like pretty big business.

Everything has a cost, and if the hiring budget is available, things like bussing several hundred interns around may be cheaper than you'd expect, certainly cheaper (even with free Xboxes) than flying hundreds of candidates out for interviews.

Using "today" prices - something like $25-50 per person for the bus, $20 per for the film, $100 for the dinner, $400 for the Xbox - less than $1k all concerned.

Thanks for putting it like this, the numbers took me by surprise. During my university years (a bit over a decade ago), I actually applied for an internship at Microsoft, and they definitely spent more than $1k flying me from Poland to UK and back for an on-site interview. I wasn't that good, my university isn't that good either, surely they could've used that money to incentivize a few local candidates with free Xboxes and come out ahead. Makes me wonder why they bothered with interns from distant lands in the first place. Tech recruitment never made sense to me (though I enjoyed benefiting from its peculiarities).

I'm 16 and I want to grow up and revive the concept of a fixed-function graphics card to make graphics cards cheaper and more focused on gaming. I might even look into acquiring the 3dfx brand and reviving it with that goal. AI and crypto whatever might be the buzzwords but they don't interest me.

GPUs are already relatively fixed-function though - they're already designed with only graphics goals in mind, which is why they're so efficient at them. The reason why they were so suitable for AI was kind of a happy coincidence, because the way they work as processors and their high memory bandwidth (both required for good graphics performance) was also perfect for AI.

“An array of programmable CPUs with instruction sets tailored towards vector operations” might be specialised but it isn’t fixed function. “Insert vertices, indices, and 2-4 textures. Receive image on screen.” DX7 style is fixed function.

Hence "relatively". Specialized is the more succinct way of describing it though, yes.

I think it is because the geeks "won". A lot of geeks found each other during the early web days. Nowadays it's not a stigma to be a geek. Everything is more mature and commercially integrated, including mopping up geek-ly minded people into commercial ventures. It's harder today to have an outsized impact from a bedroom and no budget. (Not in absolute numbers, how often this happens, but in relative terms! Many more people are at it nowadays, so some slip through with a run-away success from nothing.)

So, in other words, people today are late.

Not at all. As someone who lived through that: there was much less enthusiasm and acceptance around starting a business. It was seen as harder and the Internet was perceived as less of a sure thing.

You could argue there was more low-hanging fruit in terms of opportunity, but to anyone who was a product of the age, they weren't primed to see it. So the absolute amount of opportunity wasn't substantially different than it was before. Our mind plays tricks on us when we look into the past: "If I had known X, I could've done that." But, you didn't. And business is rife with survivor bias.

Yes and no!

Today one is probably late to

"make a random open source project and see it become a pillar of industry",

- but today is the perfect time for

"start a software business and make great money"

In the terms of too late to invent the wheel, sure, we can opine about not being that person all day and you'll always be correct. Unfortunately for you, there's more than that which makes the world we live in today and tomorrow.

> I wonder if this established leader position somehow could make people less optimistic and willing to explore?

The market seems like a bad fit for pro-humanity technology—researchers at least have that (albeit fading) potential in their back pocket.

Reply on news.ycombinator.com