@damienriehl.bsky.social
Lawyer, technologist, educator, musician, speaker. Litigator for ~15 years, coding 1985+, web 1995+, also cybersecurity. Now AI-backed tools @Fastcase and @vLex
What's the value of privacy? It's $4!
09.11.2025 09:40 — 👍 0 🔁 0 💬 0 📌 0
a16z: "The 'best' companies have hostages, not customers." Over the decades, how has this applied to #legaltech?
youtu.be/Kw2bK_jLz8M?...
In defense of the em dash — punctuation making us more human!
www.nytimes.com/2025/09/18/m...
Maybe there's room in legal tech for both #determinism and #probabilism.
Link to Thinking Machines' paper:
thinkingmachines.ai/blog/defeati...
The best lawyers I know are creative. They eschew the "most likely next token" — favoring out-of-distribution legal approaches and arguments. Some legal tasks have the "right answer"; many tasks, though, have "an answer so creative that it just might work."
23.09.2025 12:44 — 👍 0 🔁 0 💬 1 📌 0Ex-OpenAI CTO Mira Murati's company Thinking Machines is making LLMs more #deterministic.
Open questions: Do we want *all* of our #legaltech to be deterministic? Or is "legal creativity" #probabilsm a feature, not a bug? 🤔
Perplexity extrapolation of Claude data → percentages in Claude's Job Explorer:
www.perplexity.ai/search/how-m...
Claude's "Explore by Job" data:
www.anthropic.com/economic-ind...
Some notes and questions:
- That's just one model (Claude)
- How many "legal profession" queries in (1) Gemini/ Google, (2) ChatGPT, (3) Grok, (4) CoPilot?
#LegalConsumer queries? E.g., #Client queries? #WouldBeClients didn't hire lawyers because LLM output = "good enough"
#Lawyers, #judges, #arbitrators, and #mediators use #Claude. Queries:
- #Lawyers: 0.21% = 201,600 daily (6,048,000 monthly)
- #ALJs: 0.19% = 5.4M/mo
- #Arbitrators + #Mediators: 0.10% = 2.8M/mo
- #Paralegals + #LAAs: 0.05% = 1.4M/mo
- #Judges + #Magistrates: 0.03% = 28k/mo
Law firm leader on stage just now: "We've equipped each of our lawyers with the $20/month ChatGPT, telling them 'Use it for everything you can.'" How many firms are doing this? How many just don't speak about their "off the shelf #LLM use" publicly?
16.09.2025 18:34 — 👍 3 🔁 0 💬 0 📌 0
LexisNexis study: "Among these consumers who are aware of [ #LLMs ], a stunning 48% have already used some form of generative AI tool to obtain #LegalAdvice or assistance with a legal question." Wow — 48%! Oh, by the way, this was from May 2023. 🤯
www.lexisnexis.com/community/in...
Google: AI's environmental impact appears overestimated
AI QUERIES: 0.24 Wh vs earlier estimates (3-7 Wh) — like watching TV for <9 seconds
TRUE ENERGY: far lower than claims (0.24 vs 3-7 Wh)
OLD STUDIES missed cooling/infrastructure
33x efficiency (44x carbon ↓ )
youtu.be/aarDw3sooYE?...
Visions of the future with AI for people, the courts, and lawyers. I'm very excited for this article from Washington State Bar Association. With thanks to @damienriehl.bsky.social, Jayne Reardon, and J. Scott Schlegel who I have been bouncing ideas off for a year or so wabarnews.org/2025/07/16/t...
30.07.2025 23:41 — 👍 1 🔁 1 💬 0 📌 0Does an AI disclosure rule help any more than Rule 11 (or its state equivalents)? "Everything above my signature is true and correct" — whether created by AI, my first year associate, or my paralegal. AI "pinky swears" don't change behavior.
20.07.2025 17:02 — 👍 1 🔁 0 💬 1 📌 0
Microsoft: AI beats human physicians in diagnoses (Humans = 2/10, Machines = 8/10)
www.theguardian.com/technology/2...
"It looks like you're filing a motion to dismiss. Would you like some help?"
30.06.2025 22:14 — 👍 3 🔁 0 💬 0 📌 0Exactly: 🎯 . Exciting times! 🚀
30.06.2025 21:29 — 👍 4 🔁 0 💬 1 📌 0
Lovely London event with Reena SenGupta of #RSGI at the @financialtimes.com FT Innovative Lawyers event!
24.06.2025 08:39 — 👍 0 🔁 0 💬 0 📌 0Interesting — I had a nice lunch with someone from the Center for Humane Technology, which helped with this case. Thanks for passing this along!
23.06.2025 18:06 — 👍 0 🔁 0 💬 0 📌 0
...or on your favorite podcatcher:
kite.link/damien_riehl
Great fun chatting #music + #Copyright
• Copyright protects expressions, not ideas
• AI extracting uncopyrightable #ideas (vectors), not #expressions (jettisoned)?
• #MLC style royalties statute for #AI ingested works?
youtu.be/1UjXTRw2Yck?...
Dark day, here in Minnesota. Democrats shot and killed: How will POTUS respond? 🤷
kstp.com/kstp-news/to...
Dark days, here in Minnesota. 🙁 These were Democrats: How will POTUS respond? 🤷
14.06.2025 15:06 — 👍 1 🔁 0 💬 1 📌 0
RIP Brian Wilson
youtu.be/9a7TENaM7mE
Legal tech "learning" through redlines? This interview with Cursor engineers discusses doing reinforcement learning (RL) on PRs (pull requests), to allow the LLM to "learn" like a Jr. coder might.
Equally applicable to lawyers? Maybe RL on Track Changes redlines? 🤔
youtu.be/sLaxGAL_Pl0?...
Estonia's FutureLaw Conference was wonderful — filled with old friends and new ones. Tallinn is lovely, and its people are great hosts. Last year was a joy, and this year might have even been better. Thanks to Team #FutureLaw25 for an excellent meeting of smart legal minds!
02.06.2025 20:44 — 👍 1 🔁 0 💬 0 📌 0
Poor people, veterans, seniors, and children rely on LSC to fund #LegalAid organizations nationwide.
The White House budget would eliminate the LSC — harming 5M low-income families. Civil legal aid helps returns $7 for every $1 invested.
Support LSC here:
americanbar.quorum.us/campaign/457...
Hey, M&A lawyers: CEO said he used #LLM for #M&A deal, saving him $100k. Harbinger of the future? Or today? If you're not using LLMs, will this client stop hiring you? Or will you not know why the phone didn't ring? See 32:37 mark.
youtu.be/JMYQmGfTltY?...
Yeah, this is an important one. I had lunch with with someone who helped Plaintiffs craft the complaint's arguments.
IRONY ALERT: Notebook LM podcast about the order; note the AI podcasters discussing anthropomorphism!
notebooklm.google.com/notebook/9d2...
