1. The low hanging fruit is definitely drying up. There aren't a lot of opportunities to (for example) replace a human doing a job that can easily be handled by a computer at this point.
2. The remaining opportunities are a lot less "sexy." Walmart is a good example of this. Supply chain optimization is absolutely something that software helps with; intelligent order routing, demand forecasts, etc.
3. Less "sexy" doesn't mean "provides less value" IMO, even if that's not reflected in stock prices.
4. "Core operational competence" is super broad, as are supply chain improvement as a general category. Do you have thoughts on what sorts of improvements are being made here that unlock massive value?
1. People didn't actually use it to buy stuff because they want to comparison shop.
2. The devices were sold at a huge loss.
What I think has changed is that Amazon now has a lot more "products" to buy and devices that make the shopping easier. If you can ask Alexa to "order X things from the Whole Foods nearby, but prefer brands I've shopped in the past" and then you're able to confirm the order on a screen, then have it delivered to your house within a few hours, that's a much more compelling offering.
A CSV of this data is free to download per the pricing page, whereas the raw data (not sure what that looks like versus the CSV) requires a paid account.
So I guess it depends if you consider the CSV as fundamentally different from the raw data in a way that makes this clickbait.
Is your issue actually with "creating an account," or is it with "giving up your personal information?" Because as the commenter indicated, you can create an account on this site without giving up any personal information.
So while you're "correct" in the sense that you do need an account, it seems that the meat of your point (giving up personal data) has been addressed.
I think the approaches usually fail because commercial farming with chemical fertilizer is pretty efficient and optimized to work at a large scale.
A lot of agtech startups (think vertical farming) just don't get this. They don't understand where the inefficiencies are, or how their approach loses efficiency at commercial-farming scale, or how the approach doesn't integrate with other necessary machinery, etc.
Vertical farming might be having trouble but a very closely related cousin, the Dutch greenhouse, is doing fantastically. The differences are subtle because Dutch greenhouses have been used to vertically farm crops as well (though not separated in the same way that vertical farms are) and use artificial lighting. The subtle differences matter and that's why iteration in the space is so good.
The surface level discussion in this thread griping over Ellison's approach is probably the least interesting aspect of agtech to discuss but is the one that most people have an opinion on and generates the most engagement which is why this thread is filled with it.
The article specifically says that he had tech CEO's running the project.
The mistakes they made also seem pretty fundamental to farming; things like:
1. they didn't consider that a greenhouse designed for the desert wouldn't work in Hawaii
2. solar panels need to be installed differently depending on their location and never see the theoretical power generation in practice
3. immature/mature plants were growm right next to each other in a way that spread pests
4. they bought marijuana greenhouses without considering that it is grown so differently from other standard crops.
This is pretty basic stuff that should have been caught by someone with knowledge of agriculture. This seems to indicate that while they had a really smart team, they made the mistake of assuming that general AI/robotics would map 1:1 to the problems of agriculture.
The success rate of university research should have been the ultimate warning sign that you shouldn't dump half a billion into "solving agriculture." Progress in established fields like agriculture is expensive, time consuming, and (usually) incremental.
The thought that you could do all of that at once and outcompete an approach that has been refined for thousands of years is wild. Kudos to them for dreaming big, but I just don't see why they thought they had an edge here outside of "AI can solve any problem" hubris.
Based on what I know about one of the states in question, I'm thinking that "Active and Passed" means they have both a passed bill and an active bill that isn't passed. Though I'd think they'd call that "Passed and Current" to match their other nomenclature.
Which measure are you going off there? Because I see the "Chance a poor student has to become a rich adult" metric at 41% for UC Riverside and 58% for Harvard.
There is the "overall mobility" metric that favors UC Riverside, but the way that's being measured would seem to skew in favor of whichever college has students in lower quintiles (a top quintile kid can't move up 2 quintiles).
Ah, you're right, I misread. But 41% vs 58% isn't a big enough difference to pay "any amount" for IMO - and the gap is much smaller with other public universities like Irvine (55%) and SUNY Binghampton (54%).
Quintiles is a poor measure given the extremes of inequality. The Pareto distribution of income and wealth has folks in the top 10/1/.1/.01 percentiles with vastly different lifestyles compared to the other percentiles.
"Any amount of money" when the statistics are still a coin toss sounds like a gambling addict. That's insanely bad odds for "life savings" amounts of money...
What do you see as the goal of the daily ranking (and by extension the new "hot" ranking)?
I ask because personally, the daily ranking allows me to evaluate myself on an "objective" basis versus other players on a given day (i.e. if I'm 50th out of 100, I'm roughly average).
Hot rankings don't exactly give me this, because how/where I rank will depend not only on my skill, but also how long ago I made the score, the timing of other high scores, etc.
The main benefits seems to be a larger leaderboards and a higher ranking when you get a particular score. Personally, I don't see those as huge benefits.
In the style of Texas Hold'em, both players could secretly choose two seats each and write them down. Then, you'd collectively choose 5 seats to be the shared "cards."
This next suggestion would stretch the "poker" definition somewhat, but I think it retains the same characteristics (imperfect information, shared "cards").
You start from a shared list of attributes (coat color, presence of a hat, etc.) and designate a row of seats. Each person gets one attribute secretly. You wager after each stop following poker conventions.
Only downside to this is that unlike poker, your hand can get worse after a stop.
This misses the information your secret hand gives you about the other player's chances. In poker if I got an Ace in my hand the chance of the opponent having an ace lowers.
If I got somebody wearing a black coat, this has no impact on the chance of my opponent having a black coated passenger.
Discard any hands where you pick the same seats, at the end.
You can also play actual poker like this. Each player writes down random numbers 0-51 in predefined order. You reveal some of those numbers to the other player who adds their own number mod 52 to get their private hand. You all slowly reveal both numbers for shared cards. If one of your cards matches the shared card you have to start over, same deal if at the end multiple players end up with the same cards.
It’s a slow process, but when the goal is wasting time and you don’t have cards it’s a poor substitute.
1. The low hanging fruit is definitely drying up. There aren't a lot of opportunities to (for example) replace a human doing a job that can easily be handled by a computer at this point.
2. The remaining opportunities are a lot less "sexy." Walmart is a good example of this. Supply chain optimization is absolutely something that software helps with; intelligent order routing, demand forecasts, etc.
3. Less "sexy" doesn't mean "provides less value" IMO, even if that's not reflected in stock prices.
4. "Core operational competence" is super broad, as are supply chain improvement as a general category. Do you have thoughts on what sorts of improvements are being made here that unlock massive value?
reply