I love this! You've convinced me (again) I need to spend more time thinking about agriculture and tech, as should people in the following categories: a) those who eat, and b) those who want to go to Mars.
Such an interesting piece, Rhishi (as always!) Thank you for writing and sharing this. You may remember that early in 2025, we spoke about some of the work I’m leading at Giga Futures on orchestrating more integrated solutions in agriculture. This made me think of our conversation!
Reading this, I keep coming back to two parallels. First, indoor lettuce production, which has effectively “designed out” many of these harvesting constraints through tightly controlled, end‑to‑end optimization (even if it’s still a small share of total production today). Second, commodity row crops, where genetics, planting architectures, and machinery have co‑evolved over decades to make large‑scale mechanization possible.
By contrast, in crops like lettuce, breeding and system design have been largely optimized for consumer and retail traits (appearance, shelf life, eating quality) and the cheap labour market, rather than for ease of harvesting in an automated era.
So your piece really raises, for me, a forward‑looking question: as we think about robotics, should we be trying only to fit machines into today’s paradigms, or also to build new ones that are more integrated and orchestrated, from genetics through to harvest and packing?
And if that could be a goal, in such a fragmented industry, who can realistically play that “integration authority” role - e.g. at the farm enterprise level, within a specific value chain, or via new platform players that sit across breeders, equipment manufacturers, and growers? What would the practical unlocks be (standards, incentives, shared infrastructure)?
Thanks again for a great story — it really underscores how much we may need to rethink how we innovate and not just what technologies we build, if we want automation to truly work for crops like lettuce.
Yes, you are absolutely right, Christine. You do see efforts to try to change some farming practices to fit the machines. There are also efforts to genetically modify strawberries to improve harvestability. For example, with lettuce, growers adjust their bed width and try to keep it level to make it easier to harvest with a machine. There are efforts to genetically modify some crops so their stems are slightly longer, making it easier for a machine to cut lettuce without damaging the produce. These efforts take time, though operational changes can be done sooner and can be effective as well.
I love this! You've convinced me (again) I need to spend more time thinking about agriculture and tech, as should people in the following categories: a) those who eat, and b) those who want to go to Mars.
Thank you, Hollis! This was a fun one to write about.
Such an interesting piece, Rhishi (as always!) Thank you for writing and sharing this. You may remember that early in 2025, we spoke about some of the work I’m leading at Giga Futures on orchestrating more integrated solutions in agriculture. This made me think of our conversation!
Reading this, I keep coming back to two parallels. First, indoor lettuce production, which has effectively “designed out” many of these harvesting constraints through tightly controlled, end‑to‑end optimization (even if it’s still a small share of total production today). Second, commodity row crops, where genetics, planting architectures, and machinery have co‑evolved over decades to make large‑scale mechanization possible.
By contrast, in crops like lettuce, breeding and system design have been largely optimized for consumer and retail traits (appearance, shelf life, eating quality) and the cheap labour market, rather than for ease of harvesting in an automated era.
So your piece really raises, for me, a forward‑looking question: as we think about robotics, should we be trying only to fit machines into today’s paradigms, or also to build new ones that are more integrated and orchestrated, from genetics through to harvest and packing?
And if that could be a goal, in such a fragmented industry, who can realistically play that “integration authority” role - e.g. at the farm enterprise level, within a specific value chain, or via new platform players that sit across breeders, equipment manufacturers, and growers? What would the practical unlocks be (standards, incentives, shared infrastructure)?
Thanks again for a great story — it really underscores how much we may need to rethink how we innovate and not just what technologies we build, if we want automation to truly work for crops like lettuce.
Yes, you are absolutely right, Christine. You do see efforts to try to change some farming practices to fit the machines. There are also efforts to genetically modify strawberries to improve harvestability. For example, with lettuce, growers adjust their bed width and try to keep it level to make it easier to harvest with a machine. There are efforts to genetically modify some crops so their stems are slightly longer, making it easier for a machine to cut lettuce without damaging the produce. These efforts take time, though operational changes can be done sooner and can be effective as well.