TL;DR
We harvested eggplant, peppers, sunflowers, tomatoes and all throughout, the mosquitoes harvested us. Despite that, it was a good day for both farm and my tech process.
Harvesting and Weeding, Seeding and Planting
All of our work in the spring and early summer is paying off in bountifully harvests now, with mornings usually full of harvesting this and that from the fields. In particular, a full blue bin of onions takes some significant effort by two people to carry to and from the truck. A bin of eggplants can, surprisingly, also be heavy enough to require two people to carry.
When we’re not doing that, we’re weeding with either/both hoes and hands. The crops in landscaping fabric are usually quick with a team to blow through, as are the early beds of winter squash planted weeks ago. Hoeing around delicate crops like squash stems takes a keen eye and a steady hand, but can be accelerated with a person doing hand-weeding around the plant roots so that hoeing can be done more quickly.
The Mosquitoes Are Bad
Maybe because of the rain this past weekend, that the farm is on the Ipswich marshes, that there was not much of a breeze today, that there’s a new cycle of hatching, the mosquitoes were bad today. Even fully sprayed with repellant, at times they were so thick that breathing was difficult without taking some in.
The good news though is that the greenhead flies seem to be gone after the new moon (tidal changes). Their bite is like a sting whereas you might not even notice a mosquito is biting you until it’s had it’s fill. I think I walked away today with only a dozen or so bites that I can feel, so sleeping tonight shouldn’t be too hard.
What Kinds of Repellent Are Best for Organic Farming?
For skin repellent, sprays based on neem and essential oils are least-synthetic. Castor and clove oil are also options. Smells like thyme, mint, rosemary and citronella all also deter many insects that may be attracted to you.
I’ve never found a purely natural option that seems to hold up in the thick of greenhead and mosquito season. Though I have a very hard time linking to a government site (because you don’t know these days when an administration may simply erase its published information), the EPA currently has published information about DEET-based repellents.
Unfortunately in New England, we also have an abundance of ticks which carry Lyme disease, and for which standard repellents (even with DEET) simply don’t work. There is the process of impregnating your clothes with Permethrin, but that also kills beneficial insects and pollinators such as bees, so that’s simply not an option for organic farming.
Not to sound like a tough guy, but practically speaking, sometimes you just have to bear with the bugs and take the middle way if you want to sleep the night. For this, I use DEET-based repellents in the heavy season and natural repellents in the fall if anything is needed at all.
Are Well Prepared Soil Beds like Well-trained LLMs?
For a while now, I’ve known that eventually I’ll run into situations where the generic off-the-shelf LLMs (like Qwen3, Llama3, Phi4, etc.) will not be able to handle the specificity to organic farming needed for the precision and complexity of tasks I’ll be facing. Real-time prioritization, incorporation of short and long term weather forecasts, and the ability to understand and respond to the nuances of organic farming practices are all areas where specialized models and systems will be needed.
Before spending tens of thousands of dollars on the cloud compute to train my own model from scratch…something I would never do anyway…I think I’ll start by exploring other, simpler options such as Custom Embeddings. Even Langchain (which I use the community version of for the more complicated graphs of work in some of my prototypes) has a custom embeddings feature, so that’s promising.
Like a soil bed where there are some key steps that have to be done before planting, my homelab prototypes and services have been built to be a well-prepared environment for the more elaborate major research steps to take place in, such as custom embeddings and non-GPU model optimization. Though I’m only a team of one right now, I’m building this out to be a scalable system that can handle the complexity of organic farming research and development.
AI Summary from Field Notes
For today, without any fancy embeddings, this is what my AI-assisted summarization process produced:
Summary
Consolidated Summary
Main Themes
- Crop Management: Harvesting and planting of various crops (eggplant, peppers, sunflowers, tomatoes, carrots, etc.).
- Pest and Environmental Challenges: Mosquito infestation affecting planting, and seasonal challenges like chicken molting.
- Labor and Tools: Use of tools (stirrup hoes, old seeder) and manual labor for weeding, crop processing, and animal care.
- Growth and Maintenance: Monitoring tomato trellises, onion peeling efficiency, and animal health.
Key Activities
- Harvesting: Eggplant, peppers, sunflowers, and cherry tomatoes.
- Planting: Carrots, turnips, and radishes using a mechanical seeder.
- Weeding: Winter squash using stirrup hoes and hand weeding.
- Crop Processing: Onion peeling, sorting, and binning (with trash separation).
- Animal Care: Checking on chickens and pigs, providing water, and monitoring molting.
- Barn Maintenance: Cleaning and shutting down the barn.
New Things Encountered
- Tools/Equipment:
- An old seeder (possibly less efficient).
- Stirrup hoes (specific weeding tool).
- Crops/Processes:
- Bee/honeybee-sized tomatoes and heirloom beefsteak tomatoes (specific varieties).
- Onion cleaning process with bin-based sorting (trash/finished product separation).
- Observations:
- Time distortion during onion peeling (1.75 hours for 4 people).
- Chicken molting as a seasonal challenge.
Questions for Future Research
- Tool Efficiency: How effective is the old seeder compared to newer models? Can stirrup hoes be optimized?
- Pest Control: How to mitigate mosquito infestation?
- Plant Growth: How to manage tomato trellises to prevent them from “getting out of control.”
- Onion Processing: Can onion peeling be streamlined (e.g., time management, labor distribution)?
- Animal Health: How does chicken molting affect productivity or health?
- Crop Varieties: Are bee/honeybee-sized tomatoes a unique variety or a misnomer?
Suggested Actions
- Pest Control: Implement mosquito traps or barriers to protect planting areas.
- Tool Maintenance: Repair or replace the old seeder for better performance.
- Weeding Optimization: Train staff on efficient stirrup hoe use or explore alternative methods.
- Onion Processing: Introduce a rotating team system to reduce peeling time.
- Monitor Chicken Health: Track molting chickens for signs of stress or disease.
- Document Tomato Growth: Record trellis growth rates to identify management strategies.
This summary encapsulates the intern’s daily tasks, challenges, and areas for improvement, structured for clarity and actionable insights.
Part 1
Main Themes:
- Crop Management: Harvesting and planting of various crops (eggplant, peppers, cherry tomatoes, carrots, turnips, radishes).
- Pest and Environmental Challenges: Mosquito infestation, challenges in weeding, and the impact of weather/seasonal changes.
- Mechanical and Manual Labor: Use of machinery (seeder, stirrup hoes) and manual tasks (onion peeling, animal checks).
- Growth and Maintenance: Observations on plant growth (tomato trellises) and animal health (chicken molting).
Activities Performed by the Intern:
- Harvesting:
- Eggplant, peppers, sunflowers, cherry tomatoes.
- Planting:
- Carrots, turnips, radishes using a mechanical seeder.
- Weeding:
- Winter squash using stirrup hoes and hand weeding.
- Crop Processing:
- Onion peeling, sorting, and binning (including trash separation).
- Animal Care:
- Checking on chickens and pigs, bringing water, and monitoring molting.
- Barn Maintenance:
- Cleaning, shutting down the barn.
New Things Not Yet Encountered:
- Tools/Equipment:
- Use of an old seeder (possibly less efficient than newer models).
- Stirrup hoes (a specific weeding tool).
- Florida weave trellising (a method for tomato support).
- Crops/Processes:
- Bee/honeybee-sized tomatoes (a specific variety).
- Heirloom beefsteak tomatoes (a type of tomato).
- Onion cleaning process with bin-based sorting (trash/finished product separation).
- Observations:
- The time distortion during onion peeling (1.75 hours for 4 people).
- Chicken molting as a seasonal challenge.
Questions and Future Research Areas:
- Efficiency of Tools:
- How effective is the old seeder compared to newer models?
- Can stirrup hoes be optimized for faster weeding?
- Pest Management:
- How to mitigate mosquito infestation (impact on planting).
- Plant Growth Control:
- How to manage tomato trellises to prevent them from “getting out of control.”
- Onion Processing:
- Can onion peeling be streamlined (e.g., time management, labor distribution).
- Animal Health:
- How does chicken molting affect productivity or health?
- Crop Varieties:
- Are bee/honeybee-sized tomatoes a unique variety or a misnomer?
Suggested Actions:
- Pest Control:
- Implement mosquito traps or barriers to protect planting areas.
- Tool Maintenance:
- Inspect and repair the old seeder for better performance.
- Weeding Optimization:
- Train staff on efficient use of stirrup hoes or explore alternative weeding methods.
- Onion Processing:
- Introduce a rotating team system to reduce time spent on onion peeling.
- Monitor Chicken Health:
- Track molting chickens for signs of stress or disease.
- Document Tomato Growth:
- Record growth rates of trellised tomatoes to identify management strategies.
This breakdown highlights the intern’s daily tasks, new challenges, and areas for improvement or further investigation.
[end of post]
Enjoy Reading This Article?
Here are some more articles you might like to read next: