See the entire conversation

1/ Just had lunch with one of the engineering leads at a top 2 tech giant.
112 replies and sub-replies as of Jun 03 2017

2/ He observed that he sees a lot more innovation within the company than outside in the startup world (says VCs are partially to blame).
3/ Got me thinking that while the cloud has materially lowered the cost of starting a co, there are several forces holding innovation back.
4/ To start, the large platform companies (monopolies) are extracting a huge tax out of any startup trying to build a business.
5/ Because they are monopolies with extraordinary margins and profitability, the tech giants are able to hire and retain the top talent.
6/ The combination of platform tax + talent cost and shortage makes it extremely difficult for startups to compete.
7/ In Enterprise land, we rode the SaaS wave for the past 17 years and disrupted all incumbents with the same application but in the cloud.
8/ The problem is that game is largely played out. The incumbents are all really good companies in the big traditional SW categories.
9/ Founders and investors are having to look for opportunities in niche markets and verticals not yet occupied. Tough!
10/ So at least in Enterprise land, we are all looking for the next wave that will pave the path for the next 15 years of disruption.
11/ We all think ML/AI will open the door for intelligent applications. I agree. But, try to hire a top notch ML person from Facebook!
12/ I am optimistic that one of these small ripples will become the next wave, but surely it seems it won’t get easier or cheaper.
We forget that innovation doesn't have to mean the most advanced tech. FB, SNAP, TWTR, AMZN all started as simple products addressing needs.
In all four cases, the innovation was the founders understanding a gap in the market based on their own experiences and capitalizing on it.
What's harder now is that the most obvious issues have already been addressed.
In your opinion what are the small ripples? Also, for non-enterprise how can they adapt? What areas to compete on where they are weak?
I guess that last question was silly. That's the gap we as founders have to identify.
The scary part then is, where is the defensibility even after you find a gap. Snap was copied in months to curb their growth.
In your opinion, is there room for more vertical saas solutions in a single industry, esp if regulated (e.g. healthcare, energy)?
nvm, just re read and saw #9 :)
Sounds like u need to get into the real estate vertical. Industry is 5-10 years behind everything else. Plenty of opportunity.
think you're missing that large incumbents can fund huge costs of developing next level of innovation; even Magic Leap can't
for this next gen, individual talents still need a deep bench of funding, research and other talents few startups can deliver
Agreed and why I am at google.
1/ Disagree VCs partially to blame. New tech incumbents break convention: They don't wane, they get stronger w/ web/mobille network effects
2/ As SaaS crowds, my focus has shifted to API-driven models (Stedi), hidden markets (government) & industrial applications (Boeing)
3/ Issue w/ API: biz model tiered, not smooth; w/ govt tech, slower adoption (but changing*); w/ industrial, strategic value > financial $
4/ personal bets are 1, industrial giants going thru massive change, will dwarf retail reshuffling, and new software co's will be scooped up
5/ and 2, govt tech market is open, seeded a bunch two years ago & now going to strong Series As much faster than i expected.
6/ and just realized we have lunch tomorrow, so good fodder for lunch! see you tomorrow.
Looks like we have a lot to talk about. See you maniana.
I disagree as well, but people say I am biased :)
I think you should disagree if you have that position.
Time to start looking at deep tech. Bioscience, energy, things that require a PhD to do. This is where the next wave of innovation will be.
I am building a profitable non-ML/AI startup in construction that would go exponential w/ proper ML/AI dev work in our niche. You are right
Contrarian take: Most ML/AI methods will be widely known and open source will democratize methods, but not to data which bigcos will keep
gotta find hard to reach datasets (healthcare, legacy industry, etc.) or create new ones I guess (genomics, equipment, etc.)
Agree. But, I don't believe generic ML/AI tools or platforms will be sufficient, so talent will be required to apply them to use case.
Talent in ML/AI is “overrated”, at least for now. Majority of engineers try one of 12 models and see which one works.
That’s why you have to look for companies that can gather and retain larger data silos than the incumbents. Plenty out there
I thought this was the conventional wisdom?
yeah not sure what makes this a contrarian take?
Sounds better. And tweeting is a speed game.
As a VC you see so many pitches for proprietary AI and dig in for diligence and don't find it
Not if you see a large volume of pitches
hire the best ML person that doesn't want to leave their family to move to SF. Thinking SF-only means you compete against those titans.
Good thread. FWIW, I worry a lot more about incumbents' data advantage than their comp advantages WRT AI. Startups offer things they can't.
I find Microsoft and IBM's formerly "unassailable" positions and their (relative!) irrelevance now comforting for future startups.
Pay me more then. I'm not willing to take half or less salary for a .1% lottery ticket while working 50% more with lower job security.
Great thread. Seeing similar trends in consumer land.
What about blockchain?
I'd disagree because of the number of categories that now want specialized business process tools
This is the important point. Eat the chocolate first. Why shouldn't innovation be cyclical and unevenly distributed? Alas, Joseph Alois!
do you think the benefit of a start up being able to use whatever tech they want (where incumbents must use their own) outweighs this tax?
disagree re. platform tax. Hosting one's own still order of magnitude more expensive & all platform providers so far offer cheap rates
you and @scottew agree -- I need to better understand the issue -- got any links I can read?
I was not referring to infrastructure cost. More the appstore, google, FB, etc distribution tax.
ie paying the 30% rev rake for playing in someone else's sandbox, kk
Talent shortage is real but no easy solution, and more failure on US education system & immigration policy. Trump making it worse.
Not to mention the distribution leverage the platforms have or the eventual bill that startups get for being on a ‘partner’ platform
Talent cost: depends. Salaries from Tech giants 2x to 3x higher than what startups can offer. But startups still can gamble on options
Of course options can turn negative quick (paid AMT yet all loss) but culturally, some people will always choose startups over Hooli
Would help if preferential stock, and other plus types, existed not anymore. Same type/quality of stock for staff as for investors. Sorry.
Isn't their control of the data at least as important?
The concentration of power (data on % of population) described here is a huge worry => not only privacy, also competition law
Not true 4 superstar founders as they can make more founding than as employees. @BarbarianCap
I believe this to be true. But I also don't buy the "X is too big and is going to crush you" rhetoric that some people believe and spread.
It is not that X is too big and will crush you. It is that X will make it too expensive for you to have a seat at the table.
I'm not debating this. But, 1-it's a very big table, 2-it can be done, you just need to measure stuff they can't measure.
Seems any new cloud service is now renting time in someone else's yard now...kind of like how certain popular apps turn into features
Same was said about MS in the 90s. It's the competition for talent that seems like bigger challenge for startups.
I suspect MOST innovation ALWAYS comes from large companies. Startups are really all about the amazing, edge case innovations
And the great founder-led tech giants of today are spending a lot of time on those as well
The best startups are the ones that can take advantage of those innovations outside of the obvious/intended usecases.
what's the "VCs are partially to blame" explanation?
VCs are too shortsighted, too risk averse, too impatient, too chicken to fund true innovation...the typical stuff.
Is there any data / evidence cited for (1) VC shortsightedness or (2) Corps being bastions of innovation?
As a guy who's not in a "top 2" (but w close friends who are), & whose job consists of fostering inno outside of SV/USA, I can't say I agree
Easy for anyone (incl me) innovating to say their specific experience generalizes to their space, but wo supporting data we're just opining
Would certainly be interesting to explore a quantification of innovation more deeply, or to find someone who already has.
For now it seems like there's a ton of room outside of SV for truly impactful innovation, and investors are def awakening to that reality
I think the stock VC response is supposed to be “all these critiques are accurate, except us & the firm that joined us on our last deal.”
The only evidence I see to support it is startups often lack the ability stick to a roadmap and execute. Big companies get the top talent.
I do not buy that at all
Tech innovation is expensive/high risk w/o factoring in adoption. Corps have built in pipelines/incentives to take on risk that VCs don't.
I concede there's more nuance there than I could fit into 140 characters, but I would love to hear why I'm wrong.
Upon re-reading, you are right. By removing words, I made a sweeping/inaccurate generalization. Lots of tech inno is supported by VCs. Fact.
Yup, same feeling here. Realized when I re-read my tweet that I had accidentally changed the meaning trying to get down to 140 characters 😳
How can they tell true innovation when so much starts off as 1 little use case and ladders up?
Certain innovations solve problems that only big firms have (eg, using prop data) but generally protecting current revenue is 1st priority
Good thread. Similar challenges in consumer
Top 2 alone have a fair amount more capital (human and economic) available to them than VCs do. :)
I observed the opposite. Few executives can get 200 headcount to make a new product, but they can start a company to fill a product gap.
I would debate this. What does he mean by "innovation"? I've been in companies that lead the world in Patents, but no cust oriented products
Thx 4 sharing! I think ML/AI takes yrs of dedicated research 2 get 2 the cusp, like a mathematician.. Big Comps can & do invest the time :/
This was the best thread on startups and innovation I've read....
I have no idea what a "top 2" tech giant is.
Apple & Google? Amazon & Facebook? Microsoft & Salesforce?
Burroughs & Control Data Corp.
Plat. tax is misnomer. It was higher barrier w/o cloud. And, can't talent can be attracted via real equity, which (some) big guys can't do?
Thread echoes themes from Martin Ford's book. Innovation squelched. Capital and talent sequestered by success.