So, was the Cruise acquisition for the talent or the tech? How many other "Hard Tech" startups have had successful exits, when compared to how many have failed? How does this ratio compare to software startups?
Yes. That's the thing about acqui-hiring. As much as we like to pretend that with enough dollars you can buy the talent required to own a product-space, the fact is that in certain spaces, individual mortals can create billions of dollars of value. You just have to pick the right space and the right individual.
Do you have any data showing acqui-hiring of $1 billion or more? I've never heard of such an amount and retention after an acquisition is really, really hard. Seems crazy to spend so much only to have people leave after 1-2 years (or less depending on the company).
Maybe more plausible when it's the core talent + specific tech applicable to the problem?
You're buying both a base platform / proof of concept+ implementation and the key individuals responsible for creating it. Then training your people on the platform with the deeper knowledge from the key individuals in the 1-2 years you've got them employed.
If you can buy a breakthrough and enough knowledge transfer to get your people operational on adapting it, it's worth it even if you don't get any additional breakthroughs?
To take an example at the far end of the spectrum: in 1989 WPP (the world's largest owner of advertising agencies) bought The Ogilvy Group (a very large ad agency) for $864 million ($1.7 billion in today's dollars.) Ogilvy was nothing but people...there was no other value (except, maybe, the brand, but that's debatable.) At the time people criticized the deal, saying "all of their assets go down the elevator every night." But it turned out to be a great acquisition for WPP.
While the founders and first few employees may leave after an acquisition, preferring to work for themselves, everyone after that is an employee and, as long as they are managed and compensated well, will generally not care so much who the company's owner is.
I'm sure the tech was part of it, but my guess is that they were mostly buying it for the ready made team.
I think this is pretty similar to the Monsanto acquisition of Climate Corporation (where I worked) several years ago.
If you're a big traditional company, building an innovative technology team from scratch is really hard, even if you're willing to spend a lot of money on it. It takes a lot of time and effort, and you probably don't have the internal leaders or culture to attract top talent.
Buying a promising startup and continuing to operate it as an independent organization is a straightforward (if expensive) way to solve this bootstrapping problem. From a strategic perspective, it can make complete sense if the opportunity cost of getting a product to market a year or two late can be measured in billions of dollars. There are a lot of advantages to being first to market.
It's both. They still have a long way to go for autonomous driving, but they have a young ambitious team and their leadership seems very strong.
Relative to a company like Google, their weakness may be on the hardware side, but GM likely has access to very strong mechanical engineers, and they certainly now have the funds to go out and hire more folks.
I'm all for taking Medium thought pieces on the fundraising environment down a peg, but the problem with Hard Tech is that it requires a nontrivial amount of venture capital. You can't be ramen-profitable when your startup requires purchasing cars for testing or renting giant computing clusters for training AI models.
In relative contrast, it is trivial and risk-free to make a lean MVP for a generic Uber-for-X and send in a YC application.
You mean on AWS/Azure? Because I have to develop a framework (https://github.com/autumnai/leaf) with it, and there is no way I am going to introduce any more complexity than necessary that early.
Understood. Leaf looks interesting. Will be keeping an eye on it. Having just recently focused on machine learning I still haven't settled on a stack. I find your rational behind using Rust quite helpful. What kind of timeframe are we looking at for when this is production ready?
Really depends on what use case. There should be a nice production-ish demo out towards end of the month, but I am afraid I can't say anything less vague at the moment.
Does the rust core team have any plans towards being able to write kernels directly in Rust using the NVPTX LLVM backend? Otherwise if you're all just calling into CuDNN anyway, pretty narrow differentiation in performance (and big difference in development resources) between you, Torch, and TensorFlow.
You need to have a Titan X to produce comparable benchmarks, otherwise a lot of people won't take them seriously. Also the 6GB of a 980Ti already rule out quite a few popular networks we needed to run.
I was not really in a position to wait much when buying the card, and I didn't want to risk waiting a few weeks to get a used card I bought on a forum, which then comes out being just as expensive in living cost. You also as a startup don't generally have the luxury to risk waiting.
I don't regret the purchase, I more wanted to show the reality of a "hard tech" startup.
Curious what YC could do from their perspective to leverage their semi-unique position here as well.
The big cloud vendors are probably more geared towards the infrastructure side, but given problems (a) that "hard tech" requires non-trivial capex, (b) that many startups fail (nature of the space), & (c) that YC sees a lot of startups, then maybe there are opportunities to more efficiently handle "hard tech" needs by time-sharing common resources.
There are always going to be bespoke needs, but things like shared vehicle fleets with standardized interfaces, etc seem plausible when you've got a stable of 5 startups, with each being perfectly fine only testing their autonomous UAV / vehicle platform Tuesdays and Thursdays (or once a month) and letting someone else use it the rest of the time.
There's a company in my e.f. cohort who do most of their work on a box with 48GB of video memory. Another regularly spins up clusters to do machine learning on large video datasets to identify theft.
Most accelerators have partners that can get this stuff for cheap or free - including physical boxes.
It's a European accelerator that focuses on founders with deep tech (I seem to be the only person here without a doctorate in machine learning) and lets them form teams in the actual programme.
Some of the other people here are predicting (with massively higher certainty than doctors) which early-onset brain issue folks will develop alzheimers, making a standard simple way to run experiments in outer space (first launch in a few weeks), doing natural language interfaces for personal finance, and other crunchy stuff.
Exactly. I've worked on autonomous vehicles by myself to the point of acquiring a vehicle for development. But the vehicle's conditions were less than ideal and it became too expensive to repair. I am using parts from it to mock up some prototypes. Acquiring a reliable vehicle for development is too expensive. Mind you, I'm an individual working on this as a hobby project. A startup will have a harder time due to overhead costs.
If anyone wants to talk autonomous vehicles feel free to email me. I enjoy it. :)
Did you try using simulations? I built a rigid body car driving simulation in college using OpenGL. Now I get a lot of inbound interest from the self driving car geeks.
Yes, I did data simulations and analysis. My main interest is in vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) real time communications. Those things were simulated with a series of Python scripts, but there is nothing like real world testing. I am in the process of acquiring an SDR like HackRF to mock up some tests. But that also gets expensive when you have multiple SDRs transmitting or receiving data. Hopefully it will happen by the end of this year.
Anybody have some spare SDRs or some Lidars while we are at it?
I catch your drift, but I don't think it's necessarily true for all "hard tech" problems.
For example, putting together a proof of concept or even an MVP for a robotics product (maybe not drivable cars) is totally feasible, but taking that kind of product to market absolutely does require a huge amount of capital. The latter part of that equation is where something like YC is going to provide an immense amount of value.
The main issue with brassboard prototypes done on a shoestring budget is that they don't show well. On top of that, they carry the risk of not being able to go from brassboard to finished product. This was probably implied by Sam's comment that fundraising will still be harder for such a company.
Absolutely. My point was that something like YC (and other programs, I'm sure) is that they'd at least help. Fundraising for this kind of company is always going to be difficult, but a foot in the door with an investor that might actually understand the end goal is a pretty big win.
I'm all for taking Medium thought pieces on the fundraising environment down a peg, but the problem with Hard Tech is that it requires a nontrivial amount of venture capital.
I don't begrudge the fact profitability isn't tied to how much work went into building something, how novel the product is, or whatever. Those bloggers are great because they point out startups like to use the cachet of "solving hard problems" to try and impress me with their less-impressive products.
This blog post makes me laugh because why does it matter to a venture fund if they're "hard tech" or not? It shouldn't, should it? It's about business potential, right? If "hard tech" does that, wouldn't you just invest in it without bragging on it and hope the competition doesn't notice? Maybe Sam is making this post because kids are starting to wise up to the fact "hacker" isn't anything more than something they let you call yourself to feel cool when don't go home after 40 hours. Now the tech is "hard," so nobody asks "why am I wasting my life on this piece of shit?"
One problem is finding VC capital: majority of VCs in SV don't know that hard-tech so they don't want to invest into something they are not familiar with.
On the other hand, the problem might be also that founders which base their startup on technology gave up too easily when looking for funding. So they gave up and get cozy and well paid job at Amazon, Google, Microsoft, Oracle, etc.
And YC application process does not help here: if something is "hard tech", then YC is probably not going to fund it. Why? There are so many application and people who are very very good at hard tech are not type of founders YC would like to invest in since they do not have great communication skills. So their YC application is probably confusing, there is no clear pitch, etc. They cannot explain their start up as "Uber for X" - since it is.... complicated.
In other words, if we get few founders who got rich and get into VC business because they did some hard-tech then we will have more hard tech companies. As of now, majority VCs did some social network / uber-for-X and that is how they got into VC business.
This is generally true, most VCs are afraid of things that have strong technical risk. But the point of the post is that starting a few years ago some VCs are finally catching up to founders on this. YC, Lux Capital, AH, Khosla, and of course our firm are trying to increase funding in the space and will continue to.
At the University of Waterloo Waterloo, the Velocity Venture Fund (http://velocity.uwaterloo.ca/funding/velocity-fund/) awards an additional $10k prize to a hardware team every term to help offset the additional cost, so they could end up walking away with $35k (as opposed to $25k for a software company).
Sure, but that only serves to emphasize the difficulties faced by companies that need to do true R&D. Your typical HW startup has way more than 10k more costs than a typical SW startup.
(Yes, I know the issue is apps vs. tech, not SW vs. HW, but there are parallels)
This is the rationale of EF (said as someone who was in it). They've recognised that IP is VERY valuable again and are capitalising on it in a massive way.
Hard Tech means nontrivial problems in all varieties, but Altman is particularly "excited about AI (both general AI and narrow AI applied to specific industries, which seems like the most obvious win in all of startups right now), biotech, and energy."
(edit: okay, okay, maybe the article isn't clear.)
I didn't find it to be clear, otherwise I wouldn't have asked.
SV has its own jargon, and not being from SV, I think it's fair to want to clarify what is meant by a capitalized phrase like "Hard Tech" in case there is a deeper meaning known to those inside SV.
Fair enough. But I don't think "Hard Tech" is being used as SV jargon here. The first paragraph defines what is meant by "hard technology" mainly in the negative -- critics who say "SV isn't doing hard technology", for whatever value of "hard technology" they mean.
Yes. My reading of the headline was "hard" in the sense of "hardware" because hardware tends to require larger investments and more time--and it's often much harder to advance to a viable point incrementally. To be sure, these characteristics also mean that hardware tends to be more difficult than many types of software but the terms don't mean the same thing.
Lasers, satellites, really accurate rifles, decade-long life extensions, industrial meat production without a nervous system, advanced optics, cheaper / better sensors, swarm AI.
In contrast Facebook is largely just a couple of web forms and some Big O problems at scale.
I don't think parent meant to say that the technology behind Facebook is trivial. Just that it's trivial when contrasted with a bunch of technologies that are at best in prototype stage.
A decade-long life extension or growing meat without animals would both probably involve more "Hard Tech" style innovation than building Facebook.
Other observations are IMO less pat. E.g. designing cheaper/better sensors could be either harder or significantly easier than building a Facebook. Depending on the type of sensor and the cost improvement.
A lot of the FB tech circa 2010 seemed to be reinvented versions of things ex-Google FB staff used. They've done a lot of innovation since then though.
Really? Tomorrow night, Google's DeepMind is probably going to mop the floor with a 9-dan Go professional for the third game in a row. What's Facebook doing for fun these days?
>In contrast Facebook is largely just a couple of web forms and some Big O problems at scale.
Facebook is doing more than that. Mark recently claimed that they are working on "a world-class News Feed, and a world-class messaging product, and a world-class search product, and a world-class ad system, and invent[ing] virtual reality, and build[ing] drones".
Stuff that's non-obvious but amazing and makes you think sci-fi... DARPA does a lot of this kind of stuff.
Definitely non-trivial problems, most likely hardware, maybe materials science or something in the other science fields. Maybe new math, or solvers that take days to run on thousands of CPUs. Stuff that might take thousands of man-hours to develop.
Not sure how YC would be involved here, as this is expensive to develop, but DARPA has a funding model.
> A popular criticism of Silicon Valley, usually levied by people not building anything at all themselves
Why can or should criticism only come from [1] people that are building things themselves? What about the press (as only one example) or people who write books? What about people teaching in colleges? What about people leaving comments on HN or any other forum?
Criticism only valid if coming from someone building something themselves? Don't agree with that. [2]
[1] Which to me by the choice of words is what is meant by this statement.
[2] If that is the case companies should never solicit any feedback from their customers about their product or their business model.
>> A popular criticism of Silicon Valley, usually levied by people not building anything at all themselves
> Why can or should criticism only come from [1] people that are building things themselves? What about the press (as only one example) or people who write books? What about people teaching in colleges? What about people leaving comments on HN or any other forum?
Another interpretation is that while criticism can come from any source, those who are not building things themselves necessarily have a different perspective than those who do, and therefore their arguments should be heard in that context.
Also there are people not currently building but that have built something in the past (or assisted others who have done so). Such as sama or pg.
> have a different perspective than those who do, and therefore their arguments should be heard in that context.
That perspective could also be more valuable in some cases. Keep in mind that many of today's "disrupters" are people outside a traditional industry who actually do great things exactly because they have little experience and are not jaded. Fresh outlook not restrained by traditional thinking.
The problem that I had (if you want to call it that) was not that sama made that statement. It's the fact that I feel he implied a negative which was not warranted. Plus the use of "usually" when the truth is nobody is in a position to know the exact experience or background of someone offering criticism particularly what they have done in the past. [1] [2]
[1] There could be people even here on HN who have extremely significant backgrounds in building things that prefer to remain anonymous.
[2] I am now thinking maybe this is why PG has others read his essays. So he could simply change a sentence or insert a word and avoid criticism and excessive footnotes.
Why can or should criticism only come from people that are building things themselves?
Because the people who associate themselves with their products will bite their tongues for business reasons. If Sam pretends like the only people who worth hearing criticize from are exactly the group that won't voice their criticism, he can act like there isn't valuable criticism coming in.
Am I building something? Who knows, I use this account anonymously so I can call bullshit, not be a kissass because I'm worrying about my reputation and the reputation of my products (if they exist).
I think "the thing you are building is too trivial", as a criticism, implicitly expands to "you should be building something more important than the thing you're building". And it's correct to be skeptical of any "you should x" claim from someone who is not xing.
I just read it as a bit of defensiveness that slipped in to the opening. "People claim we're not doing real stuff but where's THEIR real stuff huh?" Agree the essay would've been better without it, but at least for me it's easy to skip over.
No. AFAICT, it's perceptual reverse-compression: ie, you look at a bad mp4 of a brick wall, and even though it's shitty, you mentally figure out the mass of pixels is a brick wall. Imagine if the computer actually re-added a crisp brick wall there.
Video compression algorithms already optimize for perception rather than objective measures like PSNR, so I still don't see the fundamental difference.
Any pointers on who might be a good VC outfit for SaaS type startup that actually needs to run their own clusters as opposed to throwing everything onto AWS?
> Any pointers on who might be a good VC outfit for SaaS type startup that actually needs to run their own clusters as opposed to throwing everything onto AWS?
Very curious, why do you think you need to run your own clusters, to the point where you think you need to seek unusually risk prone capital?
If your justification is that you are are optimizing for cost (e.g. maximizing your performance per dollar), then you must think the capex dramatically lowers opex at scale if you move from AWS to specialized hardware (say, an FPGA bare metal cluster instead of a GPU or CPU cluster in AWS). In such a case I would argue that you should just get enough paying customers first using AWS, to build up enough cash-flow that would then clearly justify the capex for scaling up and going colo. And at that point, you would not need to shop around for wild risk taking capital, the CBA could easily be explained to any VC and in fact you could probably even just get a bank loan to finance the capex if it is a clear winner in reducing costs.
The only scenario I can see where you really need to run your own hardware from day 1 is if you need to terminate some dedicated circuits for specialized data feeds (e.g. real-time market data)... In that scenario, I would recommend looking into co-locating the data appliance in a third-party data center that is physically nearby to US-EAST and can run a Direct Connection into AWS for you.
Otherwise, you may want to consider the possibility that your desire to own your own cluster may be a case of premature optimization and scaling up too early/too fast. It's a very expensive thing to do and an easy way to accidentally run out of money if you don't have client contracts signed beforehand.
We connect storage nodes via infiniband for lower latency and higher throughput. Think RDS type service where you can have significantly larger instances like say up 40TB DB running on
4 way box (soon 8 way) box.
Couldn't you get at least similar storage performance from EBS Provisioned IOPS SSDs in AWS? And maybe "scale out," instead of "scaling up," to maintain performance within AWS constraints? The advantage obviously is that you avoid the capex and if you don't achieve the sales that you foretasted then you can just rapidly scale down your AWS footprint, thus avoiding financial catastrophe. You could spring for the fancy custom bare metal hardware when you reach a "hyperscale" amount of workload where it makes sense to switch to your own hardware, like Google or Facebook. And if your customers really want the fancy Infiniband clusters right now, why not license your software and make them put up the capital to buy the hardware themselves? Just my two cents, maybe I'm missing something and your approach makes sense.
Nope you really can't that's why RDS, Azure SQL etc. have pretty bad limitations on max instance sizes. The service itself is integrated storage layer (think PostgreSQL + ElasticSearch + Event Bus) that can scale far beyond what is possible with current options.
It was always kind of frustrating to see the amount of brain power and tech going into the place I worked in Italy, www.CenterVue.com compared with typical 'web' startups. Those guys have computer vision, mechanical and electriconic engineering, and lots of software of various kinds. They're doing pretty well, but not nearly as well as plenty of way more "trivial" startups. I put trivial in quotes because it's never easy, but... you'd like to hope there's some correlation between effort and reward.
There is a correlation, but effort is a vector more than a magnitude. Lots of effort in the wrong direction gets you nowhere. Meanwhile small effort surfing on a trend can work quite well.
> A popular criticism of Silicon Valley, usually levied by people not building anything at all themselves, is that no one is working on or funding “hard technology”.
Lost all interest in saltman's point when I read this line.
It's a valid criticism, regardless if the people who are making it are "builders" or not.
It was a turn off, because it makes the author seem bitter about (what I think is) a valid criticism.
I don't understand -- you had a bad reaction to a single line in the introductory paragraph and knowingly let this turn you off to the entire post? Is that a valuable type of filter for you?
It's a bad way to respond to an argument, because it subtly undermines everyone who is critical of Silicon Valley. They either don't build things (and they're unqualified to have an opinion) or they do build things (but they're being brushed aside because they're a minority).
> A popular criticism of Silicon Valley, usually levied by people not building anything at all themselves,
I have to take issue here. I don't need to be a "builder" to see that what a given startup is building is either worthless or an attempt at solving a non-problem.
I'm not a helicopter pilot, but if I see a helicopter stuck in a tree, I don't need to be one to know that the guy screwed up.
Sure, but a lot of people like to complain that snapchat/tinder/etc. are wastes of time or poor uses of talent when really they make a lot of people's lives a lot better.
It's far from an objective observation, I'm merely pointing out that the idea that you need to be doing a thing in order to criticize someone else's thing is hogwash.
"I find this product useless" is useful. "Other people are stupid for finding this product useful" is useless, just as is "other people are stupid for not finding this product useful".
I worked in industrial automation for a few years and most of the software there is terribly outdated. It's definitely ripe for disruption. There are "firm tech" business ideas that involve building better sensing and control software for hard tech industries. Given the low cost of sensors and networking these days it's not prohibitively expensive.
I think making a hard tech product for consumers is still very hard, and those who succeed should be commended. People wanting to dip their toes in the water might consider things like building a factory management portal like Splunk that aggregates sensor data, or something that involves sensing and reporting rather than actuation. Actuation can be dangerous and therefore legally risky.
I think the reason these systems haven't been challenged yet is lack of knowledge. People tend to attack areas and industries that they know a lot about. This is why college hackathons have tons of dating, what to do, and delivery apps. It's their world.
If you don't have people coming out of the industrial automation space with a good understanding, it's hard to walk back in with a useful product.
Exactly. Industrial automation also isn't sexy and manufacturing margins are small so the big dogs in the industry have limited R&D resources.
The upside is a lack of competition and a B2B environment. No need for a marketing department when you just have to attend a few industry trade shows a year.
Be careful with "lack of competition". You are ALWAYS competing with the incumbent solution. No matter how clumsy or awkward it seems; no matter how much better it could be; those solutions have amazing staying power and your niche can become your enemy because there's a limit universe of customers to sell the solution to.
To expand on that, lack of awareness of the problem space and lack of knowledge of the specifics involved. And I think the latter is a huge barrier.
From what I've seen (relatively minimal), industrial automation tends to be incredibly bespoke. E.g. we had consultant X from Y integrator (sometimes now defunct) come in Z years ago (where Z is always > 10) and set up this system: we've been using it as a black box without change since.
In short the "Linux on the desktop" problem - the burden of having to support 1,000+ unique configurations, each with their own edge case behaviors (or outright bugs). And from what I've seen, the ideal startup growth pattern doesn't fit with a services "send one engineer out to an account to custom fix it" way of doing things.
"we had consultant X from Y integrator (sometimes now defunct) come in Z years ago (where Z is always > 10) and set up this system: we've been using it as a black box without change since"
This is accurate and it stems from the proprietary technologies that integrators must sew together. Want to build a better Siemens Step 7 IDE? Get a job at Siemens.
I think ultimately it stems from lack of source code ownership / use of open platforms / clearly defined interfaces.
Best case, with clearly documented and defined interfaces, you can rewrite or swap out an entire component. Worst case, you have no source, no standard technologies, and no interface barriers, in which case your choices are leave as-is (no additional features) or replace everything (impractical due to size of codebase / legacy functionality coverage).
That helps but you have to understand that most industrial and utilities sites hate new tech. They like proven tech that's been used in three other places before they adopt anything. And once they settle on something they want to stick with it for 15-30 years and have support for longer because everyone knows that it's going to take another 3 years to get approval for purchase of new components because the old stuff has been giving faulty readings the past year and a half which requires hard reboots, the firmware is no longer being upgraded and the guys who first used the stuff are starting to retire. Even then they just want something that fits into the same slot.
Sorry for the runon but this is a major gripe of mine. Most managers just want it to work and unless it doubles your profit, cuts failures and makes you coffee for when other stuff breaks and you're working overtime to fix it they don't want to change anything.
Well, naturally. Their job is not to use tech. It's to build something, or run something, etc.
I can sympathize, and I bet you can too. My dishwasher, my car, my garage door... I just want these things to do their job and let me get on with my actual work. I don't want the latest garage door or vacuum.
I find that kind of odd. Cooking is not a passion of mine, so i don't care about a fancy dishwasher, i agree i want something that just works. I'm not much of a gamer, but i completely understand friends who dump hundreds or thousands a year into hardware, because that is something they're passionate about.
If the whole point of the organization is to build things, it seems like they should use the best tools for building things. Now, there's a good argument for tried and true solutions, and if the new stuff isn't rock solid, you're going to have wasted materials. It really seems like they should be completely geeking out over better, shinier, fancier tools for building stuff.
Maybe the costs are so high, and the margins so thin there's no room at all for experimentation. But it can't be lack of desire. If it is indeed lack of desire, they're going to go out of business.
Agree and also Investors have a bias towards the student crowd as they are much easier to impress and deal with, than an experienced Academic or Engineer.
> I think the reason these systems haven't been challenged yet is lack of knowledge.
As a software engineer, I always thought that the best way to have a great business (not a blow out VC style business, but a good business) is to be a decent-to-good software developer and have deep domain knowledge of some other industry. I thought the best way to get that would be to apprentice in that industry, but the opportunity cost is high (esp as you get older), so the second best way may be to take a job as a dev in that industry and learn as much as you can from the domain experts.
The combo of software development + domain expertise means you see problems to automate away all around you and you can actually execute on what you see.
> I think the reason these systems haven't been challenged yet is lack of knowledge
As a software engineer that works in this space, I disagree. The thing that keeps us using things like MODBUS and CAN are because the problems we are solving are themselves are very simple. So we tend to use simple solutions to address them. The latest dynamic languages and protocols (XML, JSON, YAML, etc) don't solve our existing problems any better.
Quite frankly, the extreme level of complexity in deploying something like Android to control a basic IoT device is beyond appalling. 640k may not be enough for everyone, but it is more than plenty to control IoT devices.
I've actually wanted to do this for a while. I worked in large-scale analytics (Crittercism) and have a friend who does robotics work in Chicago.
I see two big problems here.
The first is, what's the business case for the buyer? What tangible improvement in their business is possible with all this new technology? Maybe less downtime? Predictive failure models? I feel like the tech is so disruptive, you'd be better off building a completely new factory from scratch rather than convincing an existing one to adopt such a radical process/ops change.
Second, there is a profound skepticism for large-scale software projects in true "industry". My father worked for a long time at Andrew, a maker of coaxial cable in Illinois, which was later acquired by Commscope; their view of software projects was a lot like war in 1984: a constant, ongoing struggle, with no end in sight, of huge promises and lack of good delivery. I'm not sure whether that means the barrier to entry is huge, or once you make great software you'll have a massive differentiated advantage -- probably both -- but just understand, these guys have spent their career dealing with vendors selling overpriced junk from Oracle, SAP, etc. and just expect things will fail from the getgo.
1) I assume you're referring to "Splunk for factories". You could make the same claim about plain-old Splunk. One business case is better real-time response to failure and later, better forensics. I agree, there are many factories where the process is perfected and the operators can read the tea leaves when something goes wrong. Another business case, and probably a better one, might be report generation for managers - people hate making the same reports over and over again.
2) Industrial software largely sucks now. But software in general sucked 20 years ago, and has improved much since then. True industry is behind the software curve by 10-20 years, so they are using old, bad (by today's standards) software. The thing is we've already improved that old, bad software - they just haven't adopted it yet. Once industry begins adopting "newer" software (think Java, .NET, web services) and they see the falling development costs and better features, they get behind it.
There is also a talent shortage in industrial software, because there are higher wages and better working conditions in the desktop, web, and mobile industries. However what has happened is we are now in diminishing returns in those latter three industries because they are saturated with talent, while all the cool things we've built to service those industries can now be repurposed by hard tech. It's the reverse of what happened after the Cold War ended and defense industry veterans fled for greener pastures in Silicon Valley.
the problem with those industries is that the software is closed because of the hardware. while on one hand it is ready for disruption, on the other it is locked down by ever changing hardware interfaces outside of your control.
You would also have a hard time convincing the buyers to not use the manufacturer software and instead use yours.
>> building a factory management portal like Splunk that aggregates sensor data, or something that involves sensing and reporting rather than actuation
Haven't ThingWorx solved this problem very well already ?
Yes. Industrial automation is very ready to be disrupted.
I'd like to see better support for actuation in terms of being able to audit things and feed that back to design decisions, which ties very much back to aggregation.
I'm biased to think there may not be much difference between the two domains. And it should be pretty clear that machine learning is going to be increasingly important.
This being said, the customers may not be all that ready for it.
Sure, but I don't consider most of those life altering events.
The new safety features on my current car, like collision detection and dynamic cruise control, have certainly changed the way I drive and so altered my life in a small way.
No need to devalue the work of others. It's likely many of these new hard tech problems will require software. Software that runs on the web. Software that is styled via CSS.
CSS styles fine, it could be a lot better in terms of ease of use for the developer.
It often feels like black magic getting something to style the way I want it, and its not something that I can sit down and work out with a pen and paper, instead it seems to rely on knowing lots of tricks and obscure features.
Building hard core technology products is very risky and challenging for entrepreneurs. Complicating matters further, recent financial market volatility and sentiments being expressed in "Medium thought pieces about when the stock market is going to crash" are amping up the fear and uncertainty faced by newcomers.
It is therefore extremely encouraging and important to hear this message clearly and directly from Sam, that this is indeed a lucrative area worth pursuing and that there are investors who have the risk appetite to continue to fund such ventures.
If I dont know anything about developing AI specifically - but have an idea of how AI can help an industry, how could I go about making that hard idea something with merit/velocity/validity?
Sure - but I'd rather bounce off someone who already knows it and can tell me if I am nuts, first. I know its a hard problem, the idea - I just dont know if its worth doing...
If it's worth doing or not is entirely detached from the methods used to do it. Your business brain should be able to figure that out. If you want to work with AI/Machine Learning do a course to understand at least the basics. Besides the fact that you won't regret it, understanding the capabilities of a technology through how it works will help you understand how to apply it. AI isn't necessarily obvious like a CRUD app is.
Thank you for this input. May you please point me in the right direction to start learning?
I am pretty senior role in my tech career and have not too much time - so even if its just a youtube channel that could start me learning by watching, I would appreciate it.
I found Andrew Ng's Coursera course to be very quick to get up to speed on the basics. If you're already in a tech role you will find it easy but informative. Definitely something you could do in a couple of evenings a week.
Sounds like a desperate scramble to rebalance the portfolio which, reputationally, is far too long of last decade's, now-stalling, strategies. Problem is the brand is not 'hard tech', it's irredeemably soft tech. The brand is consumer biz disruption and social. YC is a short (if only it were listed) just like the rest of the web stack, for reasons of acute oversupply. It is woefully ill positioned on hard technologies. Kudos for trying, though! Heroic effort.
Anyone want to start a ternary computing startup? In the post Moore world I think there is a market and it is one of those innovations none of the big players are working...
Big ass problem: Found one, where a
the first good solution is a
must have for nearly everyone
in the world with Internet access.
"Hard tech" solution: A bunch of
applied math, with advanced prerequisites,
some original, for
a unique, really good, fun to use,
interactive and
addictive,
by far the best in the
world solution
-- Did that.
"Hard"? Silicon Valley has
more hen's teeth than entrepreneurs
who could understand the theorems
and proofs of my math even if
I explained it to them. Why?
They didn't take the prerequisite
pure/applied math courses in
grad school. Neither did
more than a tiny fraction
of computer science profs.
Code: 80,000 lines of typing,
running, in alpha test -- did that.
"We hope to hear from you."
You did, and you ignored it.
Using my HN UID,
look up my submission in your records.
If you are interested now,
then let me know.
I invited Sam to contact me, and
from my HN UID his records will
give him my e-mail address so that
he can contact me.
Sam requested that anyone with "hard
tech" contact him. Here I'm taking
"hard" as difficult, challenging, or
technically advanced and not necessarily
'hard' as in 'hardware'.
I do intend to announce, say, my
beta test on HN, but for the current
alpha test that seems a bit early.
Sam asked people to contact him,
and here I did and reminded him
of my earlier submission that
he might be interested in now.
Here's a several billion dollar hard tech idea that would benefit many, many, many people: cheap/fast/efficient water desalination and purification. It's a hard problem. Current solutions aren't cutting it.
Unlock water from the oceans, and a lot of people can be helped. I'm ready to start this company, just need a few dozen engineers/chemists and a boatload of funding.
As an amateur inventor with a serious backlog of "hard tech" type inventions (e.g. actual products, not software or 'platforms') I've done my best to inquire and reach out to several different potential funding sources, all of which have been extremely quick to pass. The basic reasoning is "this doesn't fit with our preferred verticals" and I'm left to figure out what that meant other than We only speak app as a self-imagined consolation.
This article is not exactly re-assuring because the noted interests - AI, biotech, and energy - are all extremely crowded fields more than likely competing with large institutions, research organizations, and let's be honest, probably a TON of regulatory hurdles to consider if planning to do business in the US.
For context, one of my primary inventions is a mobility / utility device that would have residential, commerical, and industrial applications. By design it can be applied to a variety of uses. I've done basic patent research and the pathway looks extremely good, quite tempting for me to just go ahead and file ASAP. It's hard tech. It's not glamorous, but it's a huge market opportunity on prima face.
I've got a co-worker pal who finally got his beverage and branding invention patented and now he's in the marketing to local schools, catalogs, and if early indications are correct, he may have an genuinely lucrative future with it. In my view, he's a more likely investor target than anybody out on the West Coast or in a VC room. I simply base this perspective on personal experience and how this post sounds promising yet concludes with a short list of massive goals. Might as well wrap it up with "We just want to invest in a better mousetrap" considering the scope. Cottage industry isn't runaway freight train profit creation, I get that, but I'd also counter that if one wants to get into the next frothy bubble of biotech then good luck with that.
But YC is in the business of hopping aboard every rocket ship it can find. Sam & Co do not want to invest in small business. They want massive payouts. It's the only way the VC game works.
If you're dedicated to getting funding, you have to keep knocking on tons of doors until you find an investor that matches your particular payout goals.
The West Coast / SV venture capital scene makes a ton of noise about massive successes to entice many otherwise-profitable small business ideas to try and become huge. When they don't reach a hyperbolic trajectory, it's a VC failure, but the same business could have easily done well for the founders and early employees as a lifestyle business. A stable orbit if you will, but not going to the Moon.
Find investors who want reasonable returns from your business at your scale. Don't think that the SV VC approach is the only way to get your business idea funded.
Oh I understand, but in this case making such a public post about "come on, join us, we can do it" and then targeting essentially stud 'just graduated, have other backing' type teams in massive fields is conflicting to me. I'm basically under the assumption that any serious venture in AI, biotech, or energy can actually compete for investments outside of the YC sphere. As in, YC actually has more competition than these teams have to sacrifice to get started. It is common knowledge capital is sloshing around everywhere looking for yield and returns.
I do understand VC might not be the proper target in the purest sense but it's a similar story for most small businesses if I'm not mistaken. Like in hard bench science research - funding goes to those already being successful. Is it really practical to "boot strap" a hard tech company? I draw a lot of inspiration from the Wright Brothers, but I'm still pragmatic.
If there's a 'ground floor' arena where cottage industry would be best served, I'm all ears and would love links, guidance and direction to make progress. Not kidding, it's already been a few years and I've got another half a lifetime to keep plodding away.
The problems you described with these fields are known as "barriers to entry". They are bad for competition and foster lower-order market structures and/or price distortion in virtually any industry, not just those in the tech sector. Thus they're generally frowned upon by all but the most extreme laissez faire capitalists:
I'm in sort of the same train and am getting slow but steady progress. If you want to talk about things (without giving away anything) feel free to email me.
Have you looked into SBIR[1] grants for developing your product? You might have to do a little leg work to set yourself up to receive Federal funding but if you'd like to commercialize any technology you'll probably have to do that anyway...
Well written and mostly agreeable but I've got a lowbrow dismissal incoming: I disagree that Cruise could be considered hard tech. The hard tech in that space was done by universities competing for DARPA awards, and Google after that. Cruise took inaccessible hard tech and made it more accessible.
And that isn't an insult to Cruise. The problem space is still difficult, and they created more value tackling that than the vast majority of startups do on much easier problems. I happen to think that the opportunities for making the cutting edge more accessible vastly outweigh the opportunities for cutting new edges, and nobody should be discouraged from doing it because it isn't as glamorous as the research on the ground floor.
Interestingly, Kyle (founder of Cruise) was intimately involved with the DARPA Grand Challenge at MIT when the goal was to engineer a self-driving car.
I don't think you're being dismissive at all - if anything I think you've added some much relevant context to the discussion. I appreciate you mentioning this relationship between development and monetization.
"Cruise took inaccessible hard tech and made it more accessible."
This seems to be a standard path for businesses in the tech space. The value many of these companies bring is not the tech itself but the ability to get that tech to the masses (which is no small feat).
It's why the classic techie response to a tech product is "who cares I can already do that with {insert X inaccessible technologies glued together with shell scripts}." The original dropbox thread on HN is a great example of this.
Also here's the original Cruise HN thread where some skepticism was also expressed. Interestingly the criticisms were more that it was too hard and a $10k bolt on box wouldn't drive safely enough https://news.ycombinator.com/item?id=7933045
I'm wondering how much of that is due to how "hard" the tech is (the research had been done by Sebastian Thrun/Carnegie Mellon in the 90s) vs. the maker renaissance spurred by the drastically falling cost of hardware?
Though I suppose "true" fully autonomous self-driving cars that can handle things like rain, human driver non-verbal communication (like eye contact and the finger), and left turns onto oncoming traffic are still at least a decade away, at least according to MIT's John Leonard, the Cruise founder's likely protégé. [1]
It goes without saying that California with its perfect weather and perfect highways is a rather unfair testing ground for autonomous vehicles, as unfortunate as it is the hotspot for a lot of the autonomous driving research. (Maybe that's why the researchers more grounded in reality are the ones in Pittsburgh and MIT/Boston?)
I remember when a LIDAR unit was a $30,000 SICK laser unit, but now you have the same kind of sensor in your $100 Kinect.
In his book Let My People Go Surfing, Yvon Chouinard makes a distinction between invention and innovation. Innovation is the application of an invention to solve a problem and create a marketable product.
Chouinard is one of the most innovative folks in the history of the climbing industry, having introduced a lot of now-standard technologies like curved ice axe picks, polyester long underwear, and polyester fleece for warmth. I highly recommend the book.
Cruise Kyle Vogt's third acquired company. It's still ambiguous as to whether your first startup should be "hard tech" or not. I think in the case of self-driving cars it makes sense because automobiles are a trillion dollar market and the anxiety to stay competitive with the rapidly changing landscape is intense.
Plenty of room in this market yet. What is the biggest competitive advantage and who has it? Data & Tesla. Tesla is the only company with the appropriate data to make self-driving a reality faster than anyone else (tell me when Google starts collecting data in Norway). Data collection equipment is relatively inexpensive. Maybe figure out how to get people to attach sonar or lidar or tap into their car's computer to access that data. What else is a problem, even for Tesla? How about predicting pedestrian behavior and making eye contact -- or identifying if a pedestrian is disabled and sending a signal to those pedestrians that it is OK to walk.
At the end of the day the market being targeted has huge weight to the likelihood of a successful outcome.
Good post. Congrats to Cruise for the acquisition and to YC for generating a return on their investment. There are definitely opportunities in AI-narrow applications using machine learning-and biotechnology right now.
This acquisition fits exactly with GM's plans outlined by investing $500 million in Lyft to create self-driving car infrastructure [0].
It seems like companies who build products that have clear and significant business value with excellent market timing are successful. Perhaps you can extrapolate to saying hard tech is now a hot field but this is the first successful exit that comes to mind recently, rather than private investors putting more money on the table.
301 comments
[ 2.3 ms ] story [ 356 ms ] threadIt has to be the tech or tech + talent. Are there enough people worth $1 billion or more to do only a talent acquisition?
Do you have any data showing acqui-hiring of $1 billion or more? I've never heard of such an amount and retention after an acquisition is really, really hard. Seems crazy to spend so much only to have people leave after 1-2 years (or less depending on the company).
Uber, Apple, Tesla, and every car company would be interested.
You're buying both a base platform / proof of concept+ implementation and the key individuals responsible for creating it. Then training your people on the platform with the deeper knowledge from the key individuals in the 1-2 years you've got them employed.
If you can buy a breakthrough and enough knowledge transfer to get your people operational on adapting it, it's worth it even if you don't get any additional breakthroughs?
While the founders and first few employees may leave after an acquisition, preferring to work for themselves, everyone after that is an employee and, as long as they are managed and compensated well, will generally not care so much who the company's owner is.
Although the "technology" more or less worked, it was discarded and re-written or substantially changed.
In the case of Google, I'm thinking of AdMob.
In the case of Yahoo, I'm thinking of the 2-sided ad exchange network they bought and then ripped out the 2-sided features.
I think this is pretty similar to the Monsanto acquisition of Climate Corporation (where I worked) several years ago.
If you're a big traditional company, building an innovative technology team from scratch is really hard, even if you're willing to spend a lot of money on it. It takes a lot of time and effort, and you probably don't have the internal leaders or culture to attract top talent.
Buying a promising startup and continuing to operate it as an independent organization is a straightforward (if expensive) way to solve this bootstrapping problem. From a strategic perspective, it can make complete sense if the opportunity cost of getting a product to market a year or two late can be measured in billions of dollars. There are a lot of advantages to being first to market.
Relative to a company like Google, their weakness may be on the hardware side, but GM likely has access to very strong mechanical engineers, and they certainly now have the funds to go out and hire more folks.
In relative contrast, it is trivial and risk-free to make a lean MVP for a generic Uber-for-X and send in a YC application.
10's of Teraflops is cheap.
Source: I just did that.
You can get the 980 Ti w/ 6GB of ram for $600 and some change. And if that's too much, the 970 w/ 4GB can be found for under $400.
And if even that is too high, you can troll gamer forums for used cards and cut that cost in 1/2.
You need to have a Titan X to produce comparable benchmarks, otherwise a lot of people won't take them seriously. Also the 6GB of a 980Ti already rule out quite a few popular networks we needed to run.
I was not really in a position to wait much when buying the card, and I didn't want to risk waiting a few weeks to get a used card I bought on a forum, which then comes out being just as expensive in living cost. You also as a startup don't generally have the luxury to risk waiting.
I don't regret the purchase, I more wanted to show the reality of a "hard tech" startup.
3.5GB usefully. The last 512MB performs significantly slower.
The big cloud vendors are probably more geared towards the infrastructure side, but given problems (a) that "hard tech" requires non-trivial capex, (b) that many startups fail (nature of the space), & (c) that YC sees a lot of startups, then maybe there are opportunities to more efficiently handle "hard tech" needs by time-sharing common resources.
There are always going to be bespoke needs, but things like shared vehicle fleets with standardized interfaces, etc seem plausible when you've got a stable of 5 startups, with each being perfectly fine only testing their autonomous UAV / vehicle platform Tuesdays and Thursdays (or once a month) and letting someone else use it the rest of the time.
Most accelerators have partners that can get this stuff for cheap or free - including physical boxes.
Probably just lots of traffic to blame
Some of the other people here are predicting (with massively higher certainty than doctors) which early-onset brain issue folks will develop alzheimers, making a standard simple way to run experiments in outer space (first launch in a few weeks), doing natural language interfaces for personal finance, and other crunchy stuff.
[0] - http://aws.amazon.com/activate/
If anyone wants to talk autonomous vehicles feel free to email me. I enjoy it. :)
Anybody have some spare SDRs or some Lidars while we are at it?
I don't begrudge the fact profitability isn't tied to how much work went into building something, how novel the product is, or whatever. Those bloggers are great because they point out startups like to use the cachet of "solving hard problems" to try and impress me with their less-impressive products.
This blog post makes me laugh because why does it matter to a venture fund if they're "hard tech" or not? It shouldn't, should it? It's about business potential, right? If "hard tech" does that, wouldn't you just invest in it without bragging on it and hope the competition doesn't notice? Maybe Sam is making this post because kids are starting to wise up to the fact "hacker" isn't anything more than something they let you call yourself to feel cool when don't go home after 40 hours. Now the tech is "hard," so nobody asks "why am I wasting my life on this piece of shit?"
On the other hand, the problem might be also that founders which base their startup on technology gave up too easily when looking for funding. So they gave up and get cozy and well paid job at Amazon, Google, Microsoft, Oracle, etc.
And YC application process does not help here: if something is "hard tech", then YC is probably not going to fund it. Why? There are so many application and people who are very very good at hard tech are not type of founders YC would like to invest in since they do not have great communication skills. So their YC application is probably confusing, there is no clear pitch, etc. They cannot explain their start up as "Uber for X" - since it is.... complicated.
In other words, if we get few founders who got rich and get into VC business because they did some hard-tech then we will have more hard tech companies. As of now, majority VCs did some social network / uber-for-X and that is how they got into VC business.
(Yes, I know the issue is apps vs. tech, not SW vs. HW, but there are parallels)
(edit: okay, okay, maybe the article isn't clear.)
SV has its own jargon, and not being from SV, I think it's fair to want to clarify what is meant by a capitalized phrase like "Hard Tech" in case there is a deeper meaning known to those inside SV.
In contrast Facebook is largely just a couple of web forms and some Big O problems at scale.
Yes, and Steph Curry is just a guy shooting a ball in a basket.
A decade-long life extension or growing meat without animals would both probably involve more "Hard Tech" style innovation than building Facebook.
Other observations are IMO less pat. E.g. designing cheaper/better sensors could be either harder or significantly easier than building a Facebook. Depending on the type of sensor and the cost improvement.
Facebook is doing more than that. Mark recently claimed that they are working on "a world-class News Feed, and a world-class messaging product, and a world-class search product, and a world-class ad system, and invent[ing] virtual reality, and build[ing] drones".
Based on the glaring flaws in the base product, I'm unconvinced it has the smarts to do any of those things well.
Stuff that's non-obvious but amazing and makes you think sci-fi... DARPA does a lot of this kind of stuff.
Definitely non-trivial problems, most likely hardware, maybe materials science or something in the other science fields. Maybe new math, or solvers that take days to run on thousands of CPUs. Stuff that might take thousands of man-hours to develop.
Not sure how YC would be involved here, as this is expensive to develop, but DARPA has a funding model.
Why can or should criticism only come from [1] people that are building things themselves? What about the press (as only one example) or people who write books? What about people teaching in colleges? What about people leaving comments on HN or any other forum?
Criticism only valid if coming from someone building something themselves? Don't agree with that. [2]
[1] Which to me by the choice of words is what is meant by this statement.
[2] If that is the case companies should never solicit any feedback from their customers about their product or their business model.
> Why can or should criticism only come from [1] people that are building things themselves? What about the press (as only one example) or people who write books? What about people teaching in colleges? What about people leaving comments on HN or any other forum?
Another interpretation is that while criticism can come from any source, those who are not building things themselves necessarily have a different perspective than those who do, and therefore their arguments should be heard in that context.
Also there are people not currently building but that have built something in the past (or assisted others who have done so). Such as sama or pg.
> have a different perspective than those who do, and therefore their arguments should be heard in that context.
That perspective could also be more valuable in some cases. Keep in mind that many of today's "disrupters" are people outside a traditional industry who actually do great things exactly because they have little experience and are not jaded. Fresh outlook not restrained by traditional thinking.
The problem that I had (if you want to call it that) was not that sama made that statement. It's the fact that I feel he implied a negative which was not warranted. Plus the use of "usually" when the truth is nobody is in a position to know the exact experience or background of someone offering criticism particularly what they have done in the past. [1] [2]
[1] There could be people even here on HN who have extremely significant backgrounds in building things that prefer to remain anonymous.
[2] I am now thinking maybe this is why PG has others read his essays. So he could simply change a sentence or insert a word and avoid criticism and excessive footnotes.
Because the people who associate themselves with their products will bite their tongues for business reasons. If Sam pretends like the only people who worth hearing criticize from are exactly the group that won't voice their criticism, he can act like there isn't valuable criticism coming in.
Am I building something? Who knows, I use this account anonymously so I can call bullshit, not be a kissass because I'm worrying about my reputation and the reputation of my products (if they exist).
Hah! Exactly. Look at my footnote [1] to my comment in this thread.
Magic Pony (machine learning for reverse-compressing video data, re-adding data that was removed during the compression process) is pretty cool.
Very curious, why do you think you need to run your own clusters, to the point where you think you need to seek unusually risk prone capital?
If your justification is that you are are optimizing for cost (e.g. maximizing your performance per dollar), then you must think the capex dramatically lowers opex at scale if you move from AWS to specialized hardware (say, an FPGA bare metal cluster instead of a GPU or CPU cluster in AWS). In such a case I would argue that you should just get enough paying customers first using AWS, to build up enough cash-flow that would then clearly justify the capex for scaling up and going colo. And at that point, you would not need to shop around for wild risk taking capital, the CBA could easily be explained to any VC and in fact you could probably even just get a bank loan to finance the capex if it is a clear winner in reducing costs.
The only scenario I can see where you really need to run your own hardware from day 1 is if you need to terminate some dedicated circuits for specialized data feeds (e.g. real-time market data)... In that scenario, I would recommend looking into co-locating the data appliance in a third-party data center that is physically nearby to US-EAST and can run a Direct Connection into AWS for you.
Otherwise, you may want to consider the possibility that your desire to own your own cluster may be a case of premature optimization and scaling up too early/too fast. It's a very expensive thing to do and an easy way to accidentally run out of money if you don't have client contracts signed beforehand.
Lost all interest in saltman's point when I read this line.
It's a valid criticism, regardless if the people who are making it are "builders" or not.
It was a turn off, because it makes the author seem bitter about (what I think is) a valid criticism.
Time is valuable. When writing an essay, the introductory paragraph is how the author gets the reader involved in it.
That line reads to me like bitterness, so I concluded that it wasn't worth my time to read any further.
And this being the internet, I felt it was worth my time to post a comment saying how I feel.
It wasn't my intent to post a comment about the other points in the article (which I obviously did not read).
I have to take issue here. I don't need to be a "builder" to see that what a given startup is building is either worthless or an attempt at solving a non-problem.
I'm not a helicopter pilot, but if I see a helicopter stuck in a tree, I don't need to be one to know that the guy screwed up.
[1] http://zoox.co/
I think making a hard tech product for consumers is still very hard, and those who succeed should be commended. People wanting to dip their toes in the water might consider things like building a factory management portal like Splunk that aggregates sensor data, or something that involves sensing and reporting rather than actuation. Actuation can be dangerous and therefore legally risky.
If you don't have people coming out of the industrial automation space with a good understanding, it's hard to walk back in with a useful product.
The upside is a lack of competition and a B2B environment. No need for a marketing department when you just have to attend a few industry trade shows a year.
Even if FactoryTalk is absolutely horrid.
From what I've seen (relatively minimal), industrial automation tends to be incredibly bespoke. E.g. we had consultant X from Y integrator (sometimes now defunct) come in Z years ago (where Z is always > 10) and set up this system: we've been using it as a black box without change since.
In short the "Linux on the desktop" problem - the burden of having to support 1,000+ unique configurations, each with their own edge case behaviors (or outright bugs). And from what I've seen, the ideal startup growth pattern doesn't fit with a services "send one engineer out to an account to custom fix it" way of doing things.
This is accurate and it stems from the proprietary technologies that integrators must sew together. Want to build a better Siemens Step 7 IDE? Get a job at Siemens.
Best case, with clearly documented and defined interfaces, you can rewrite or swap out an entire component. Worst case, you have no source, no standard technologies, and no interface barriers, in which case your choices are leave as-is (no additional features) or replace everything (impractical due to size of codebase / legacy functionality coverage).
Sorry for the runon but this is a major gripe of mine. Most managers just want it to work and unless it doubles your profit, cuts failures and makes you coffee for when other stuff breaks and you're working overtime to fix it they don't want to change anything.
I can sympathize, and I bet you can too. My dishwasher, my car, my garage door... I just want these things to do their job and let me get on with my actual work. I don't want the latest garage door or vacuum.
If the whole point of the organization is to build things, it seems like they should use the best tools for building things. Now, there's a good argument for tried and true solutions, and if the new stuff isn't rock solid, you're going to have wasted materials. It really seems like they should be completely geeking out over better, shinier, fancier tools for building stuff.
Maybe the costs are so high, and the margins so thin there's no room at all for experimentation. But it can't be lack of desire. If it is indeed lack of desire, they're going to go out of business.
I meant identifying the problem and coming up with a solution in the first place.
Go to market is another level of complexity.
As a software engineer, I always thought that the best way to have a great business (not a blow out VC style business, but a good business) is to be a decent-to-good software developer and have deep domain knowledge of some other industry. I thought the best way to get that would be to apprentice in that industry, but the opportunity cost is high (esp as you get older), so the second best way may be to take a job as a dev in that industry and learn as much as you can from the domain experts.
The combo of software development + domain expertise means you see problems to automate away all around you and you can actually execute on what you see.
As a software engineer that works in this space, I disagree. The thing that keeps us using things like MODBUS and CAN are because the problems we are solving are themselves are very simple. So we tend to use simple solutions to address them. The latest dynamic languages and protocols (XML, JSON, YAML, etc) don't solve our existing problems any better.
Quite frankly, the extreme level of complexity in deploying something like Android to control a basic IoT device is beyond appalling. 640k may not be enough for everyone, but it is more than plenty to control IoT devices.
I see two big problems here.
The first is, what's the business case for the buyer? What tangible improvement in their business is possible with all this new technology? Maybe less downtime? Predictive failure models? I feel like the tech is so disruptive, you'd be better off building a completely new factory from scratch rather than convincing an existing one to adopt such a radical process/ops change.
Second, there is a profound skepticism for large-scale software projects in true "industry". My father worked for a long time at Andrew, a maker of coaxial cable in Illinois, which was later acquired by Commscope; their view of software projects was a lot like war in 1984: a constant, ongoing struggle, with no end in sight, of huge promises and lack of good delivery. I'm not sure whether that means the barrier to entry is huge, or once you make great software you'll have a massive differentiated advantage -- probably both -- but just understand, these guys have spent their career dealing with vendors selling overpriced junk from Oracle, SAP, etc. and just expect things will fail from the getgo.
1) I assume you're referring to "Splunk for factories". You could make the same claim about plain-old Splunk. One business case is better real-time response to failure and later, better forensics. I agree, there are many factories where the process is perfected and the operators can read the tea leaves when something goes wrong. Another business case, and probably a better one, might be report generation for managers - people hate making the same reports over and over again.
2) Industrial software largely sucks now. But software in general sucked 20 years ago, and has improved much since then. True industry is behind the software curve by 10-20 years, so they are using old, bad (by today's standards) software. The thing is we've already improved that old, bad software - they just haven't adopted it yet. Once industry begins adopting "newer" software (think Java, .NET, web services) and they see the falling development costs and better features, they get behind it.
There is also a talent shortage in industrial software, because there are higher wages and better working conditions in the desktop, web, and mobile industries. However what has happened is we are now in diminishing returns in those latter three industries because they are saturated with talent, while all the cool things we've built to service those industries can now be repurposed by hard tech. It's the reverse of what happened after the Cold War ended and defense industry veterans fled for greener pastures in Silicon Valley.
You would also have a hard time convincing the buyers to not use the manufacturer software and instead use yours.
Haven't ThingWorx solved this problem very well already ?
I'd like to see better support for actuation in terms of being able to audit things and feed that back to design decisions, which ties very much back to aggregation.
I'm biased to think there may not be much difference between the two domains. And it should be pretty clear that machine learning is going to be increasingly important.
This being said, the customers may not be all that ready for it.
The new safety features on my current car, like collision detection and dynamic cruise control, have certainly changed the way I drive and so altered my life in a small way.
It often feels like black magic getting something to style the way I want it, and its not something that I can sit down and work out with a pen and paper, instead it seems to rely on knowing lots of tricks and obscure features.
It is therefore extremely encouraging and important to hear this message clearly and directly from Sam, that this is indeed a lucrative area worth pursuing and that there are investors who have the risk appetite to continue to fund such ventures.
I am pretty senior role in my tech career and have not too much time - so even if its just a youtube channel that could start me learning by watching, I would appreciate it.
Follow up with cs224/cs231 if interested.
Big ass problem: Found one, where a the first good solution is a must have for nearly everyone in the world with Internet access.
"Hard tech" solution: A bunch of applied math, with advanced prerequisites, some original, for a unique, really good, fun to use, interactive and addictive, by far the best in the world solution -- Did that.
"Hard"? Silicon Valley has more hen's teeth than entrepreneurs who could understand the theorems and proofs of my math even if I explained it to them. Why? They didn't take the prerequisite pure/applied math courses in grad school. Neither did more than a tiny fraction of computer science profs.
Code: 80,000 lines of typing, running, in alpha test -- did that.
"We hope to hear from you."
You did, and you ignored it.
Using my HN UID, look up my submission in your records. If you are interested now, then let me know.
I invited Sam to contact me, and from my HN UID his records will give him my e-mail address so that he can contact me.
Sam requested that anyone with "hard tech" contact him. Here I'm taking "hard" as difficult, challenging, or technically advanced and not necessarily 'hard' as in 'hardware'.
I do intend to announce, say, my beta test on HN, but for the current alpha test that seems a bit early.
Sam asked people to contact him, and here I did and reminded him of my earlier submission that he might be interested in now.
At least for now, the next step is up to Sam.
Unlock water from the oceans, and a lot of people can be helped. I'm ready to start this company, just need a few dozen engineers/chemists and a boatload of funding.
That's why it's hard.
Can't make use of the science portion of it if the engineering portion isn't ready, too.
This article is not exactly re-assuring because the noted interests - AI, biotech, and energy - are all extremely crowded fields more than likely competing with large institutions, research organizations, and let's be honest, probably a TON of regulatory hurdles to consider if planning to do business in the US.
For context, one of my primary inventions is a mobility / utility device that would have residential, commerical, and industrial applications. By design it can be applied to a variety of uses. I've done basic patent research and the pathway looks extremely good, quite tempting for me to just go ahead and file ASAP. It's hard tech. It's not glamorous, but it's a huge market opportunity on prima face.
I've got a co-worker pal who finally got his beverage and branding invention patented and now he's in the marketing to local schools, catalogs, and if early indications are correct, he may have an genuinely lucrative future with it. In my view, he's a more likely investor target than anybody out on the West Coast or in a VC room. I simply base this perspective on personal experience and how this post sounds promising yet concludes with a short list of massive goals. Might as well wrap it up with "We just want to invest in a better mousetrap" considering the scope. Cottage industry isn't runaway freight train profit creation, I get that, but I'd also counter that if one wants to get into the next frothy bubble of biotech then good luck with that.
But YC is in the business of hopping aboard every rocket ship it can find. Sam & Co do not want to invest in small business. They want massive payouts. It's the only way the VC game works.
If you're dedicated to getting funding, you have to keep knocking on tons of doors until you find an investor that matches your particular payout goals.
The West Coast / SV venture capital scene makes a ton of noise about massive successes to entice many otherwise-profitable small business ideas to try and become huge. When they don't reach a hyperbolic trajectory, it's a VC failure, but the same business could have easily done well for the founders and early employees as a lifestyle business. A stable orbit if you will, but not going to the Moon.
Find investors who want reasonable returns from your business at your scale. Don't think that the SV VC approach is the only way to get your business idea funded.
I do understand VC might not be the proper target in the purest sense but it's a similar story for most small businesses if I'm not mistaken. Like in hard bench science research - funding goes to those already being successful. Is it really practical to "boot strap" a hard tech company? I draw a lot of inspiration from the Wright Brothers, but I'm still pragmatic.
If there's a 'ground floor' arena where cottage industry would be best served, I'm all ears and would love links, guidance and direction to make progress. Not kidding, it's already been a few years and I've got another half a lifetime to keep plodding away.
The problems you described with these fields are known as "barriers to entry". They are bad for competition and foster lower-order market structures and/or price distortion in virtually any industry, not just those in the tech sector. Thus they're generally frowned upon by all but the most extreme laissez faire capitalists:
https://en.wikipedia.org/wiki/Barriers_to_entry
[1]: https://www.sbir.gov/about/about-sbir
And that isn't an insult to Cruise. The problem space is still difficult, and they created more value tackling that than the vast majority of startups do on much easier problems. I happen to think that the opportunities for making the cutting edge more accessible vastly outweigh the opportunities for cutting new edges, and nobody should be discouraged from doing it because it isn't as glamorous as the research on the ground floor.
This seems to be a standard path for businesses in the tech space. The value many of these companies bring is not the tech itself but the ability to get that tech to the masses (which is no small feat).
It's why the classic techie response to a tech product is "who cares I can already do that with {insert X inaccessible technologies glued together with shell scripts}." The original dropbox thread on HN is a great example of this.
Though I suppose "true" fully autonomous self-driving cars that can handle things like rain, human driver non-verbal communication (like eye contact and the finger), and left turns onto oncoming traffic are still at least a decade away, at least according to MIT's John Leonard, the Cruise founder's likely protégé. [1]
It goes without saying that California with its perfect weather and perfect highways is a rather unfair testing ground for autonomous vehicles, as unfortunate as it is the hotspot for a lot of the autonomous driving research. (Maybe that's why the researchers more grounded in reality are the ones in Pittsburgh and MIT/Boston?)
I remember when a LIDAR unit was a $30,000 SICK laser unit, but now you have the same kind of sensor in your $100 Kinect.
[1] https://www.youtube.com/watch?v=x5CZmlaMNCs
Chouinard is one of the most innovative folks in the history of the climbing industry, having introduced a lot of now-standard technologies like curved ice axe picks, polyester long underwear, and polyester fleece for warmth. I highly recommend the book.
Cruise Kyle Vogt's third acquired company. It's still ambiguous as to whether your first startup should be "hard tech" or not. I think in the case of self-driving cars it makes sense because automobiles are a trillion dollar market and the anxiety to stay competitive with the rapidly changing landscape is intense.
Plenty of room in this market yet. What is the biggest competitive advantage and who has it? Data & Tesla. Tesla is the only company with the appropriate data to make self-driving a reality faster than anyone else (tell me when Google starts collecting data in Norway). Data collection equipment is relatively inexpensive. Maybe figure out how to get people to attach sonar or lidar or tap into their car's computer to access that data. What else is a problem, even for Tesla? How about predicting pedestrian behavior and making eye contact -- or identifying if a pedestrian is disabled and sending a signal to those pedestrians that it is OK to walk.
At the end of the day the market being targeted has huge weight to the likelihood of a successful outcome.
Incidentally, if you want to get into YC: https://twitter.com/hunterwalk/status/708304772055441408
It seems like companies who build products that have clear and significant business value with excellent market timing are successful. Perhaps you can extrapolate to saying hard tech is now a hot field but this is the first successful exit that comes to mind recently, rather than private investors putting more money on the table.
[0] http://www.reuters.com/article/us-gm-lyft-investment-idUSKBN...