DietaryNonsense

I understand there's a difference between the stated values and actual values of individuals and organizations, and so I want to ask this in the most pragmatic and consequentialist way.

I know that labs, institutions, and so on have safety teams. I know the folks doing that work are serious and earnest about that work. But at this point are these institutions merely pandering to the notion of safety with some token level of investment? In the way that a Casino might fund programs to address gambling addiction.

I'm an outsider and can only guess. Insider insight would be very appreciated.

I was inspired to ask this after seeing Farmbot mentioned on Hacker News again this morning. One of the comments more accurately describes that project as a gardening robot.

I am both a farmer and a software engineer. Like many, I have more ideas for software projects than I can realistically undertake, and my time is further limited by the demands of running my farm.

Traditional farming typically involves very high-cost, large machinery used across hundreds or thousands of acres. Gardening, on the other hand, is much smaller in scale and involves more careful tending to individual plants or smaller groups of crops, rather than massive fields like a 50-acre plot of wheat.

There is a middle ground that might be called "Small Farming" or "Market Gardening." Our production space is about 2 acres, and our beds are a standard 30 inches wide.

Over the years, I've seen many articles and posts about new AgTech products, such as AI-driven laser weeding, drone spraying, and the now common autopilot systems aboard tractors and other machinery.

What I would love to see, and may end up starting myself, is an Open Source FARM Robot. Perhaps something like a wheeled unit with a wheelbase that comfortably fits over the standard 30-inch bed. The drive unit/chassis could allow for modular payloads like a weeder, seeder, or amendment hauler and distributor (compost, fertilizer, etc.).

Is anyone aware of projects or folks working on something along these lines?

Disclaimer: I have only shallow knowledge of LLMs and machine learning algorithms and architecture in general.

Once a model has been trained, the totality of it's knowledge is presumably encoded in it's weights, architecture, hyper-parameters, and so on. The size of all of this presumably being measurable in terms of number of bits. Accepting that the total "useful information" encoded may come with caveats about how to effectively query the model, in principal it seems like we can measure the amount of useful information that's encoded and retrievable from the model.

I do sense a challenge in equating the "raw" and "useful" forms of information in this context. An English, text-only wikipedia article about "Shitake Mushrooms" may be 30kb but we could imagine that not all of that needs to be encoded in an LLM that accurately encodes the "useful information" about Shitake mushrooms. The LLM might be able to reproduce all the facts about Shitakes that the article contained but not be able to reproduce the article itself. So in some ontologically sensitive way, the LLM performs a lossy transformation during the learning and encoding process.

I'm wondering what we know about the data storage characteristics of the useful information encoded by a given model. Is there a way in which we can measure or estimate the amount of useful information encoded by a LLM? If some LLM is trained on Wikipedia, what is the relationship between the amount of useful information it can reliably reproduce versus the size of the model relative to the source material?

In the case of the model being substantially larger than the source, can I feel metaphorically justified in likening the model to being both "tables and indices"? If the model is smaller than the source, can I feel justified in wrapping the whole operation in a "this is fancy compression" metaphor?

Direct Air Capture (DAC) and related methods for CO2 capture and sequestration are often offered as an eventual solution to excess atmospheric CO2.

It seems well understood that current technology in this space is not sufficient and preliminary in terms of success and results, but it also seems to be the case that plenty of serious attention and resources are being directed into this general approach.

I'm wondering if someone can provide us with an overview / breakdown of the thermodynamics involved in this type of methodology.

There's surely a hard minimum on the amount of energy required to remove CO2 from the atmosphere (moving the air, breaking the bonds of the molecules, pumping or transporting the post-reaction products, etc). There's also the issue of the carbon emissions required to produce these facilities to begin with (even if we assume all energy inputs are derived from renewable sources).

There are many options for IP security cameras and multi-camera setups. Ubiquiti, Foscam, Nest, Ring, and all of the things. But they all involve running untrusted internet connected devices on a local network. I want to improve my physical security with these devices without providing nodes to some future DDoS botnet or whatever else these poorly secured IoT devices will get repurposed for. I also don't want my system to be useless if the internet goes down or if BigCompany decides to change their terms or drop service for their APIs or whatever else.

Wondering if anyone has had success in setting up a self-hosted (maybe open source) camera system for their site? And if so, any advice? recommendations? sources for information that you found useful?