Alex Hamilton
alexhamiltonrad.bsky.social
Alex Hamilton
@alexhamiltonrad.bsky.social
CEO and co-founder of Radiant Law, curious about what happens if lawyers stop being so... lawyerly
This account is incredible
"Use an open-source legal bill review tool to identify areas of inefficiency within your firm's billing practices. This can help you streamline processes and ensure compliance with industry standards." #LegalTechTips #FirmManagement
December 2, 2025 at 12:36 AM
Greatest ad ever
October 3, 2025 at 11:39 PM
Reposted by Alex Hamilton
Here is a PDF of chapter 17 - it's just 3 pages and will fill you with hope and optimism, or so I've been told share.eva.mpg.de/index.php/s/...
August 20, 2025 at 11:57 AM
My first interaction with ChatGPT in 2022
April 10, 2025 at 10:48 PM
Hi Claude, create a Chrome plugin for me that every 1-90 minutes during working hours (randomly) first sends a prompt to the OpenAI API that requesting text for a new prompt that looks like it relates to doing work at a law firm, then the plugin opens a new tab with url XXX....
April 4, 2025 at 11:20 AM
Well that was a surprising week with a lost laptop leading to me becoming a Mac user, and the US joining the axis of evil. Boo being mean to biglaw too, although I have a track record.

Wonder what happens next week
March 7, 2025 at 11:46 PM
Actually.... process -> assets/tools -> metrics -> process and data is either an asset (a valuable list) or the feed for metrics.

Assets/tools probably break down into platforms and contents or something
Still musing. Perhaps the cycle is process -> assets -> data -> metrics -> process...

Assets includes tools

Anyway the point about all of this is repeatability as the basis for improvement. So if you want to inject a random word generator in then you need to wrap it in controls
December 10, 2024 at 10:21 PM
I sometimes get told that Radiant is toast, because AI. I tried to model the argument...

Any issues jump out for you?
December 10, 2024 at 8:11 PM
Reposted by Alex Hamilton
I wrote a list of 15 types of work where AI is particularly helpful, and also 5 areas where it can be a trap.

AI is still best, at least for now, as a co-intelligence working with humans. Knowing when to use it, and when to avoid it, is a useful skill. www.oneusefulthing.org/p/15-times-t...
15 Times to use AI, and 5 Not to
Notes on the Practical Wisdom of AI Use
www.oneusefulthing.org
December 9, 2024 at 2:57 PM
Reposted by Alex Hamilton
@joepatrice.bsky.social delivering insanely good comedy gold as usual: ‘Grok spits out...all the humor and wit of a concussed rooster pecking on an Ouija board.’ 🐓😂 abovethelaw.com/2024/12/elon...
Elon Musk Feeds AI ‘All Court Cases,’ Promises It Will Replace Judges Because He’s An Idiot
It’ll render ‘extremely compelling legal verdicts.’ Sure….
abovethelaw.com
December 3, 2024 at 9:48 PM
Still musing. Perhaps the cycle is process -> assets -> data -> metrics -> process...

Assets includes tools

Anyway the point about all of this is repeatability as the basis for improvement. So if you want to inject a random word generator in then you need to wrap it in controls
November 27, 2024 at 6:34 PM
What about people? People are everywhere. They create/follow/improve the processes, they create the data (directly or indirectly), they hopefully review the metrics. That's where Shingo's brilliance about the focus on behaviour comes in. Work on the inputs, not the outputs.
November 20, 2024 at 11:08 PM
I'm just going to keep chatting to myself. On to continuous improvement which I think basically boils down to improving processes and tools/assets.
November 20, 2024 at 11:02 PM
Building out the process -> data -> metrics point a bit further, there are basically two aspects to continuous improvement: improving the process and improving the tools. Tools include assets, which are particularly important in knowledge work - templates, playbooks, checklists etc etc.
November 20, 2024 at 11:01 AM
There's a term in bank IT departments of "change the bank" v "run the bank". I think the process-> data-> metrics-> process -> cycle in my last tweet (? what are they called here?) is basically how does one do iterative improvement in the run state.
November 19, 2024 at 5:21 PM
I think I'm going to hang out at the other place now
November 23, 2024 at 5:17 AM
Hello world

I'm going to use this nice quiet space as a place to develop ideas.

First up is a pattern I've observed as we have built Radiant: process -> data -> metrics -> process... and round we go
November 18, 2024 at 7:19 PM
It was never going to be one-and-done: thinking evolved, CLMs and GenAI blew up, and the team kept figuring things out. The second edition is being published on 24 June, with fresh content on purpose, ideal behaviours, and why relationships matter most (plus AI!).
November 23, 2024 at 5:17 AM
v2
November 23, 2024 at 5:17 AM
Need to update our list of document automation platforms (pure plays, not CLMs) for our doc auto guide. We still love Docassemble, but what else is out there?
November 23, 2024 at 5:17 AM
Trying to get MS Copilot to do something useful and it just seems to be a nobbled version of GPT. Has anyone got any value out of it?
November 23, 2024 at 5:17 AM
Legal, sales, procurement
November 23, 2024 at 5:24 AM