FarmLogs(blog.samaltman.com) |
This has become more of an issue with farmers over the past year, as they realize their data is a commodity just like what they raise--whether crops or livestock. How much of an issue? Two examples.
First, the Iowa Power Farming show was last week. One of the largest ag shows in the nation, and the back of the program had an advertisement from Ag Leader which started with the line: "Settling for a precision farming partner that wants control of your data just doesn't cut it." Additionally, several vendors in the precision ag and sensor/drone areas were stating similar lines. Whether BS or not, they wouldn't be saying it if there wasn't a need for them to.
Second, the January 17, 2015, edition of Iowa Farmer Today also addressed this issue in their editorial page article "Who own precision technology data?".
It's not that I'm against using technology for ag. It's I'm against others using my data about my farming operations AND MY LAND and making money off of it without cutting me in--or selling/providing it (raw or as a "derivative work") to other entities I may not want to have it.
That said, it looks like we aren't doing a good enough job making it obvious in our ToS. That page has not been updated in a very long time. I am going to be sure to update it to more accurately reflect our values.
Until the investors/board say otherwise. :)
Been there. Experienced it. Resigned rather than comply.
Understand FarmLogs is the whipping boy here because you're the thread's subject (and your TOS clause 7 really is scary), but my concerns are also with the others, including for example Deere.
And as your TOS are updated, things might want to consider:
1) Who owns the data--the landowner or the tenant? Which data specifically? Who can order it removed? 2) What happens when a farm changes owners and the new owner wants all data regarding his land removed? 3) What happens when the data is subpeonaed in a clean water lawsuit for agricultural nitrate runoff?
1) Is this actually a concern to the people you want to be selling to, as evidenced by talking to a number of them? Don't bother with surveys, just straight-up have a non-sales conversation about it with 10 people who match your target customer model.
2) Lots of people don't read a ToS. Or if they do, they're not going to be 100% confident that their interpretation is what will stand up. If this is a legitimate concern for your users, you need to allay that concern, directly, in your primary sales contexts. Make it a focus item on your landing page, with design such that people will see it and go "oh, okay, I don't need to have that concern after all."
I liken it to mining. The companies who dig for various mineral deposits make significant investments into exploration of the ground. Their exploration generates a ton of data. They use that data to determine where to dig further and estimate the financial potential of the deposits. This then determines company stock values. It's ludicrous that such data may not be owned by the mining company that generated it (or acquired the rights for it). Obviously, what's happening with farms and FarmLogs is not a perfect analogy. But the parallels are interesting.
What is stopping a local server/you own the data solution? Is it the cloud solution is the only one on offer or is it too hard to get the hardware/software to the gate?
Many of these companies are also moving towards cloud/SaaS-based solutions and developing private networks to facilitate real-time data ingestion from their equipment (tractors, combines, planters, sprayers, irrigation systems, wagons, etc).
So there are local solutions, and they're installs like any other software.
http://www.wired.com/2015/02/new-high-tech-farm-equipment-ni...
Particular as a farmer I imagine you have little time to be screwing around with a local server that's prone to failure. A company looking after that for you is probably quite tempting.
I'm strongly rooting for https://sandstorm.io/ here - I hope they'll be able to provide a good alternative for SaaS businesses.
Which is funny because many VCs in SV ask the question (explicitly or implicitly) "is this something two Stanford CS students can do?" If the answer is yes, they move along.
I want more FarmLogs stories to exist.
Nice to see a new startup is trying and succeeding to make a positive difference for some pretty critical customers :)
Going to keep an eye on this group for sure. Something tells me they'll be around for a while.
The other major early infusion of tech into farming I can think of is the growth of GPS guided precision application in the early 00's. I was in college at the time and my dad worked for a farm services company; they had people that went out and drove around fields with equipment figuring out exactly where to spray the right stuff. I recall it being a big deal because it saved the farmers a lot of money on chemicals.
As an example, corn and soybeans are commonly rotated to help reduce the usage of nitrogen-based fertilizers in fields (soybeans put nitrogen in the ground, corn takes it out). If you have better data about nitrogen fixation, you can know if/when you should swap, and/or run the math on expected prices for soybeans vs. corn, irrigation costs, etc. Normally this is done in spreadsheets, but in the app you can get estimates and budgets for all of this stuff much faster (at least from what I read).
Farming is hard, and better support for data/analytics on farms is awesome.
I'm guessing a big chunk of that figure is derived from a boost in crop yield due to optimal placement. So the bottom line cost for the farmer was similar but money went much further.
Maybe something like 20% of row crop farms of acreage greater than X where X is fairly large?
I wish there were a bit more meat as to HOW they did this. Or a link to a meatier write-up of the path they took from idea to present.
It's interesting to read updates about them as they pop up.
When I saw this post I went running to the Jobs page, and there my heart got smashed when I saw Clojure on backend positions... I am a plain Java dev with learning skills (worked with Groovy/Python/JS a while ago) but I definitely don't like Clojure.
Good luck guys
Perhaps he's promoting some other JVM language -- he did mention Java and Groovy. If you type "Groovy" into HN search, you'll notice more than the usual amount of Groovy stories being submitted over the past week, and the recent comments mentioning "Groovy" are scattered around different submissions rather than clustered as usual. Groovy has more than doubled its percentage on Tiobe between January and February (if you get what I mean), and its project manager has just finished running a campaign among his Twitter followers to double its Github stars from under 600 a month ago. Because the Groovy and Grails project managers are now competing for their funding, expect to see more comments of that nature.
Lisp like syntax all those (((())))))(()() drive me crazy and I think Scala has the features of Clojure that I like without the LISPness.
But I have nothing against people using I just don't like it.
Does anyone have thoughts / strategies around identifying these opportunities, ideally without the need to spend years in the target industry?
* Your job (harder if it's just building tools for web developers)
* Your family members and friends' jobs & hobbies
* Basically anyone willing to talk about their problems with you
There are some ways to speed it up by talking to lots of experienced people, reading analyst papers, researching, or even shadowing a person in the industry for a week. However, nothing beats first hand knowledge.
A very punny one, though :)
PS hi chris.
My in-laws run a vineyard & winery. For many years, I was convinced I would build a sensor-net and on-premises app with some automation capability (toggling servos for irrigation, "fertigation" type system, tanks, timers, etc). Never really made it out of idea stage as my priorities took a different turn, but glad to see some other folks tackling the problems.
Seems like the super low cost of sensors now makes this really attractive (heck, a small farm of couple acres could be wired up to a handful of strategically placed rasberyPi's in a weatherproof case).
Sam Altman is right - this is the sort of thing that two software devs can pull off - but like everything the door must be closing.
I'm tempted to make a spreadsheet of all the novel data sources I can find (ONS, satellite, travel, weather etc) and plot against SEC codes - I am pretty sure that will allow me to procrastinate long enough for the enthusiasm to wear off.
I somehow remember the original text to be
> And importantly, they started out doing something that any two programmers could have done.
There is nothing on the wayback machine either. Am I misremembering the original text?
So the right question to ask is "what is the distribution edge of this particular team?" For example, if you're first to market and have incredibly high growth, there is a good chance it will be impossible for other teams to catch up.
Product complexity is almost never a competitive advantage, and I think most VCs understand this much better than you give them credit for.
When working in VC, it was refreshing to talk with someone not doing social/mobile/local/consumer. Once you branch out to industries like Mining, Oil&Gas, Agriculture, Automotive, Manufacturing, Logistics, Insurance, etc. just having 2 bright software engineers working on a problem at all was often enough to provide an edge.
Insurance is a good example. That industry move $2T every year in the US, which makes the US ad industry look tiny. However, it only sees maybe 1/20th the number of SV teams.
That may be true for your typical cloud / SaaS / consumer internet / AirBnb for XYZ -kind of startup but there is many many fields in technology where killer teams are exceptionally rare.
Most of these technology fields are not as sexy as building the next Dropbox but very often all the more sophisticated in their core tech. Think about bio-tech, energy, nanotech, lasers, (space) flight, AI, etc. It is never bad for a startup to assume that there is competition and to research it but it is very possible for the right people in a particular field to team up and be the best.
For example in the (UK) building trade there is a booming trade in "builder remediation" - fixing the accidental below spec work they did not notice. It can wipe out much of the profit a builder makes having to go in and pull a wall down.
I keep calling the solution to this as "bringing the factory to the field". That is measurements and control processes we see in the factory applied in less controlled environments. With mobile tech for example one could slap QR stickers on timbers, fittings and hard boards and at each stage of the build require a photo of the materials in-situ before they get covered in plaster or the next layer up.
This would help track an enormous number of issues in the building trade.
I can see how just a little inside knowledge can help.
"And importantly, they started out doing something that any two programmers (with domain expertise in their market) could have done."
and the key part of that is the "domain expertise in their market."
Our socioeconomic system very carefully prevents most students with farming domain expertise from ever getting near the Stanford CS Dept. It seems like most of the value add here is from the domain knowledge, not from the technology.
While I agree with the fact you shouldn't be screwing around with a server, that doesn't mean running your own server with your own data on it isn't a good idea. It's just we're not quite there yet!
Rich Hickey is on the record as saying that he doesn't want to promote Clojure, and if people like it they like it. He'd rather have people who came to it voluntarily than who were marketed to.
I would like to know real answers to why people don't like it. I've used a (normal) number of languages over the years but I think Clojure is by far the best for all kinds of reasons.
It's interesting to talk to people who have gone as far as trying it but found they preferred something else. I can understand you might be put off from a distance by the parens, or the immutability or whatever, but opinions change (or don't) once you actually try something.
As for what I couldn't get used to in Clojure, even after 2 years the circumfix parentheses still seemed unnatural, even though I know they're necessary for macros. I guess using f(x,y) notation in school and one's first programming language is a bond too hard to break.
I played with Scala too... Maybe I am receiving a paycheck from all other langs besides Clojure to kill the damn thing.
Come on...
I just don't like the LISP like syntax and the general idea of it and I think that other languages like can provide the same nice pieces it does... Simple... Simple... No aliens, JFK Zombie killers, Umbrela corps or anything.
So either the citation is out of step with the original page or a goofball put that in Wikipedia.
[1]: http://www.oxforddictionaries.com/us/definition/american_eng...
I'm not sure. I tend to assume the market is efficient and I found that far more often than not, the assumption proves correct. I don't know very much about most of the fields you mentioned, but think of space flight -- multiple space flight companies were founded around the same time as SpaceX (including John Carmack's now defunct space company, Virgin Galactic, Orbital Sciences, Blue Origin, Planetary Resources, and probably many others I'm missing).
I think that as a rule of thumb, the startup market is much more efficient than it originally seems, even in very deeply technical, capital intensive fields.
(defn blub-extra [a b]
(blub (inc a) (inc b)))
def blubExtra(a: Int, b: Int): Int {
blub(inc(a), inc(b))
}
Integer blubExtra(Integer a, Integer b) {
return blub(inc(a), inc(b));
}
It's roughly the same numbers of brackets (or equivalent) in Clojure, Scala and Java. They're just in slightly different places.
(excuse syntax I've not written the last two for years)
I guess my LISP Teacher ruined the syntax for me. Terrible terrible class.
Maybe you are right. I might give it a try again and see if my position still stand :D