The monetization wave is coming.

I think my personal opus token use at work is like $15K a month.

So as a billing attorney, how do you make up the income from the lost hours of billing for your work from efficiency gains?

I run a frac sand business and are main competitive advantage is proximity. I am always going to Google Maps to see the distance from one of our plants vs. our competitors to a potential job. The issue with that is you can only see one route at a time and can't compare the two.

I told Claude my issue and said I want to be able to see the distances and compare the two. It built me an incredible interface that shows the distances, how much better the ideal route is, and visualizes all this. I then got to play around with it some more and input all of my plants and my competitors as well as trucking data to show how much the customer will save. I did all this in basically no time at all. It's amazing.

tmaggie50 said:

---

So as a billing attorney, how do you make up the income from the lost hours of billing for your work from efficiency gains?

---

Take on more work.

You've probably heard the story about the two campers and the huge Grizzly Bear. One starts frantically putting on his shoes, and the other says, "why are you doing that? You can't outrun that bear!"

"I don't have to outrun the bear; I just have to outrun you."



Clients will come to expect speed and reduced cost. I aim to know more about AI than any other lawyer in my area. Adapt or die.

MavsAg said:

---

I run a frac sand business and are main competitive advantage is proximity. I am always going to Google Maps to see the distance from one of our plants vs. our competitors to a potential job. The issue with that is you can only see one route at a time and can't compare the two.

I told Claude my issue and said I want to be able to see the distances and compare the two. It built me an incredible interface that shows the distances, how much better the ideal route is, and visualizes all this. I then got to play around with it some more and input all of my plants and my competitors as well as trucking data to show how much the customer will save. I did all this in basically no time at all. It's amazing.

---

I was planning my garden at my house. I am doing new raised bed this year. I didn't ask, but it created a program with toggle switches I could see several different layouts. Once I selected the layout, I asked it to figure out how to best fill them. It gave some German word for the method. Mulch, cheap dirt, really good dirt, vermiculite, perlite. It gave me the volume of each needed for my six new beds.

I then asked it to search Home Depot, Lowes, and Tractor Supply for the best price, preferrably at one location unless there was a huge difference in cost.

It searched all the sites, even calculated upcoming sales promotions and factored in the 5% discount I had with a Lowes credit card. It even found another local provider I wasn't even aware of.

The 5% discount prevailed. It told me the number of bags of each product to get. I ordered it all online.

I drove my truck up to Lowes with my trailer and I didn't think about it, but it was loaded to capacity. I got worried so before I drove off, I took a picture of my trailer info on a sticker on the side. I asked it to calculate the weight based upon the bags and the load capacity for the trailer. It told me I was a few hundred pounds overweight, and how many bags to transfer to the bed of my truck. I went over that for safety, but that was awesome.

The more I play with it the more possibilities I see but there are things that are incorrect or incomplete or dated. When you call them out it corrects it's answers to some degree but you have to have a level of knowledge to see the flaws. I suggest people delve into AI on a subject you already know well to get a feel for the nuances.

I use it at work to help with sql code all the time. It's good a good aide but tends to over engineer what could be a simpler solution.

Tried using Gemini this past weekend for horse racing. Not for picks but breeding lines, tendencies, surface preference. It was worse than worthless. Just completely making crap up. I spent more time checking and correcting it than I did actually handicapping.

How meta. Some AI bullsh/t comes on a post about AI to pimp some stupid AI tool.

Not Chat GPT (which I use at home a lot), but my EMR just rolled out an AI product and I subscribed. I walk into the room, my MA loads up the patient's chart on the iPad and hits the record button. In the past, my MA would scribe for me and had to pretty much stay in the room the whole time or completing the note was left up to me. Now, they can walk out, get the next patient ready, clean instruments, answer patient calls, etc. I dictate my findings and plan out loud, which is better for the patient anyway, the AI wraps it all into a note and once I reconcile (there are a few hallucinations and errors, but very few), and the note is done.

I also use Open Evidence on the iPad whenever something is tricky, I need to sort through multiple options, etc. Just today I had an old lady with a really strange condition and I asked Open Evidence what meds might be causing her symptoms and what the best alternatives would be and it spat out the answer in seconds.

Especially in medicine, however, it is "Garbage in - garbage out". But so far it is a huge time saver and I really think my delivery of medicine is better for it.