Live data from Hacker News

Ask HN: Agriculture startups doing interesting work?

news.ycombinator.com

201–205 of 205 posts

Re: Ask HN: Agriculture startups doing interesting work?

#201
Hi,

I am looking for people in the image/satellite mapping industry. The idea is to map Indian farmland according to the survey no and provide the data directly to the farmer. Stack the data with the crop, soil, weather, market, buyer and it could be a solution to the whole Agri sector. If it is further mapped with land type data, Land purchase, subsidy, infra development red tape can be reduced to the extreme and the whole process could become transparent.

I have a couple of startups in mind who have the tech platform, but not the combined application stack.

I am looking for someone from a tech background to help with finalizing the whole setup. Connect me over at dishank.zala@gmail.com

Re: Ask HN: Agriculture startups doing interesting work?

#202
post #143

I worked for The Climate Corporation when it was acquired by Monsanto (now Bayer) for $1B. There is certainly plenty of potential for innovative tech in agriculture, but it’s a really tough space. Machine learning and big data are not magic bullets. The level of noise in experimental measurements is very high, which makes it difficult to prove the effectiveness of almost any intervention. Climate Corp did (and still…

Is this FieldView? Well..hold my beer! Field view collected so much data that the farmer doesn’t need(aka can’t use). And paid for it. There is so much data in the field that is worth more than the crop that is being sold. So..no..the farmer doesn’t get to monetize it. Farmers shouldn’t have to pay for any data collected that isn’t used by him but can be commoditized. In fact farmers should be PAID for allowing compa…

There is undoubtedly some value in this data, but it is far less than you think. Despite the hype about big data, it’s hard to draw firm conclusions from any uncontrolled observational trial.

Consider how often nutrition studies contradict each other. We still don’t know for sure if it’s better to eat less red meat!

Having spent a bit of time working in this space, I can assure you that it is basically the same scenario. If Climate Corp could prove that the data they are collecting with FieldView was valuable, they would be making a heck of a lot more money.

Re: Ask HN: Agriculture startups doing interesting work?

#203
post #185

I'm doing some work for a startup that wants to automate greenhouse climate control. The software automatically manages lighting, climate, irrigation based on long term growth strategies, real time sensor data and energy prices. This saves time spent manually managing greenhouse climate, and money due to improved crop and energy management. We're hiring! Looking for a data scientist with a mind for optimization probl…

You must be jeroen

Maybe I am, maybe I ain't... ;)

Re: Ask HN: Agriculture startups doing interesting work?

#204

Earlier quoted context omitted.

marketing strategy is key. The tech & science behind the yield optimization -- up to 6x 'overclocked photosynthesis' using aeroponics (not soil), optimized for photonic efficiency -- humans can accelerate the growth of plants substantially effectively "overlocking" nature. it's based on the Minskey/MIT media lab designs and some pieces I found in China -- a training gym for growing plants, but with a non-laboratory '…

> humans can accelerate the growth of plants substantially effectively "overlocking" nature. For the production of biomass you need photosynthesis. The coupling between the amount of photons you supply and the produced biomass is strong and uncicumventable. So even when you implement all the nice optimizations to accelerate plant growth, you still need to supply a proportional amount of light. If this light is artifi…

@jsilence - the GrowPotBot is micro-scale but using relays it could control a much bigger system so a single unit could control a greenhouse or space-station. Just like a PC could control one light in your home or many.

For a space station i'd recommend having at least three of them and a quorum protocol to detect failures. (planned, but not in the MVP)

now let's get into "GrowPotBot is a toy which doesn't solve any problems" .... hmm, okay let's start by ignoring the urban food deserts in poverty stricten areas and focus on the "food that looks like food but isn't really nutritious due to it being harvested before it was ripe and then sat in a refrigerated storeroom or transport ship for a year before you got it" .. so that creates a MASSIVE carbon footprint between refrigeration and storage. it was also irradiated and chemically treated to destroy any bugs before it was allowed to move across borders.

Also -- at least in the land down under (where I am) in Australia we have the largest density per capita of hydroponic growers in the world and something curious happened because .... well, FUCK capitalism where a handful of growers cut costs and used nasty fertilizers and made 'unhealthy toxic hydro vegetables' or at least that is NOW what the public perception, so a lot of people want more control over what they eat and GrowPotBot delivers that.

i personally don't like getting my food from nameless industrial agriculture who is using who knows what pesticides, if you're cool with that then good on ya!

which brings me to respond to your last point that sunlight is somehow better. umm.. no. that's false.

i love sunlight; getting the vitamin D. but sunlight is only available in limited quantities during the daytime and plants can actually grow and grow and grow for 18-20 hours per day. plants actually do marginal photosynthesis in moonlight btw. plants evolved with the sun we have, so yeah -- of course sunlight works; but they're actually quite 'inefficient'

if you've got sunlight and can control heat WHILE keeping the pests and organisms (soil fungus) away then absolutely it's possible to design an enclosure which uses natural sunlight. the bigger it gets; the more difficult a proposition that is.

there's some fascinating research in this area; but suffice to say that frequency and intensity with a intelligent optical sensors in a closed feedback loop (positronic farming) is something "new" -- and "new" ag-tech is exceptionally rare. the thing is -- it's more computer/electronics than it is farming, and farmers like to rely on natural systems versus man-made systems which is why I'm skipping the farmers and going straight to the consumers.

now for the science:

plants use red 650nm-720nm & blue !420nm and reflect green light. red produces fruit "veg" and blue produces growth "blue" and again green is reflected. this is known, google "why leaves are green" and how chlorophyll operates to generate sugar and growth.

solar panels take in full spectrum photons regardless of frequency including x-ray, gamma rays, bla bla .. and newer [future] versions actually convert waste heat back into current -- this means that we're effectively able to take in full spectrum sunlight and change the frequency to something plants can use. i suspect we'll see clear solar panels / nano-films at some point.

GrowPotBot is using 12v LED fobs doped for specific frequencies which have very (absurdly) high efficiency rates and very low heat dumps compared to traditional grow rooms which often require massive cooling. so please don't compare us to those medieval monolithic architectures with giant high pressure sodium monsters. thanks!

Post reply on HN