Jordan Furlong
@jordanfurlong.com
Legal sector analyst, consultant, author, speaker, and reformer (he/him). I publish a free bi-weekly Substack newsletter: https://jordanfurlong.substack.com/ "We have the chance to turn the pages over." Luke 12:24-32.
Wouldn’t want to be in Caracas right now.
November 13, 2025 at 12:12 AM
Wouldn’t want to be in Caracas right now.
No, no. It was someone I’m reasonably certain isn’t on BlueSky.
November 12, 2025 at 8:35 PM
No, no. It was someone I’m reasonably certain isn’t on BlueSky.
AI is just beginning its transformation of the world economy, and wealth destruction is likelier to precede wealth creation for several years to come. Don't get too attached to or invested in early AI leaders.
I'm not saying OpenAI is Netscape or Betamax. But I'm not not saying that either.
I'm not saying OpenAI is Netscape or Betamax. But I'm not not saying that either.
November 5, 2025 at 3:34 PM
AI is just beginning its transformation of the world economy, and wealth destruction is likelier to precede wealth creation for several years to come. Don't get too attached to or invested in early AI leaders.
I'm not saying OpenAI is Netscape or Betamax. But I'm not not saying that either.
I'm not saying OpenAI is Netscape or Betamax. But I'm not not saying that either.
Above all, Trump is a con man and a grifter. Everything he says and does has to be understood in that light. All the domestic and foreign chaos and grief his presidency has sown is secondary, for him, to his self-enrichment and self-aggrandizement. That's the Rosetta Stone of his presidency.
October 31, 2025 at 4:41 PM
Above all, Trump is a con man and a grifter. Everything he says and does has to be understood in that light. All the domestic and foreign chaos and grief his presidency has sown is secondary, for him, to his self-enrichment and self-aggrandizement. That's the Rosetta Stone of his presidency.
Reposted by Jordan Furlong
Polling is not an indicator of peoples' deeply-held, immutable, moral convictions — it's a function of people pressing a button on a phone after having had three seconds to think about a topic.
It is *a* data point, not a fact of nature.
It is *a* data point, not a fact of nature.
October 29, 2025 at 5:28 PM
Polling is not an indicator of peoples' deeply-held, immutable, moral convictions — it's a function of people pressing a button on a phone after having had three seconds to think about a topic.
It is *a* data point, not a fact of nature.
It is *a* data point, not a fact of nature.
If trying to bribe someone to betray their boss doesn't say "manly lethality and the warrior ethos," I don't know what does.
October 28, 2025 at 7:08 PM
If trying to bribe someone to betray their boss doesn't say "manly lethality and the warrior ethos," I don't know what does.
4? I’m Canadian and I’ve been to 16.
October 26, 2025 at 8:17 PM
4? I’m Canadian and I’ve been to 16.
Reposted by Jordan Furlong
ordinary americans are filming ice agents and trying to save their neighbors. for all the hemming and hawing during 2024 about how democracy was an elite concern, the elites gave up far more quickly than average people.
October 24, 2025 at 1:51 PM
ordinary americans are filming ice agents and trying to save their neighbors. for all the hemming and hawing during 2024 about how democracy was an elite concern, the elites gave up far more quickly than average people.
I'm inclined to the theory that his aides are showing him AI simulations of riots in Portland and telling him it's real, and he's too addled to know any better.
October 23, 2025 at 8:04 PM
I'm inclined to the theory that his aides are showing him AI simulations of riots in Portland and telling him it's real, and he's too addled to know any better.
October 22, 2025 at 3:27 PM
Just sent you a DM.
October 17, 2025 at 4:45 PM
Just sent you a DM.
I'm hoping the full list of research questions might be included in the full report if they send it. My info comes from these two articles: www.law.com/legaltechnew... and www.legaltechnologyhub.com/contents/can.... A couple of the questions are provided below:
October 17, 2025 at 4:37 PM
I'm hoping the full list of research questions might be included in the full report if they send it. My info comes from these two articles: www.law.com/legaltechnew... and www.legaltechnologyhub.com/contents/can.... A couple of the questions are provided below:
Most of the findings appear to be here: www.vals.ai/industry-rep.... I also emailed them requesting a copy of the full report, but I haven't yet heard back, and I don't know if those contents are much different than what's provided on this page.
Vals AI
Private, domain-specific benchmarks in legal, tax, and finance.
www.vals.ai
October 17, 2025 at 4:33 PM
Most of the findings appear to be here: www.vals.ai/industry-rep.... I also emailed them requesting a copy of the full report, but I haven't yet heard back, and I don't know if those contents are much different than what's provided on this page.
Bottom line: Every type of AI provider will get better at legal research. I see no signs that lawyers are getting better at it, and several signs (e.g., bsky.app/profile/giff...) they're getting worse.
Is the problem that AI is so great? Or is it (also) that lawyers aren't as good as we think? //
Is the problem that AI is so great? Or is it (also) that lawyers aren't as good as we think? //
I think one of the most problematic aspects of this is that the barrister—called in 2006 & holding an English LLB—was unable to find a copy of a Court of Appeal judgment which is available on BAILII. It raises a question of how basic legal research was carried out before generative AI became popular
This latest "barrister meets AI" disaster is something all stage 1 law students should read. It is vital about the key skill of case reading - you mustn't use AI to substitute for the skill of being able to find cases and find key content in cases:
tribunalsdecisions.service.gov.uk/utiac/ui-202...
tribunalsdecisions.service.gov.uk/utiac/ui-202...
October 17, 2025 at 3:23 PM
Bottom line: Every type of AI provider will get better at legal research. I see no signs that lawyers are getting better at it, and several signs (e.g., bsky.app/profile/giff...) they're getting worse.
Is the problem that AI is so great? Or is it (also) that lawyers aren't as good as we think? //
Is the problem that AI is so great? Or is it (also) that lawyers aren't as good as we think? //
1. ChatGPT almost matches the performance of the specialized (and much pricier) legal-specific AIs.
2. Human lawyers at the "top" US law firms finished well behind both types of AI.
3. The major legal AI platforms (LexisAI, TR Co-Counsel, and vLex) all declined to participate. I don't know why.
2. Human lawyers at the "top" US law firms finished well behind both types of AI.
3. The major legal AI platforms (LexisAI, TR Co-Counsel, and vLex) all declined to participate. I don't know why.
October 17, 2025 at 3:23 PM
1. ChatGPT almost matches the performance of the specialized (and much pricier) legal-specific AIs.
2. Human lawyers at the "top" US law firms finished well behind both types of AI.
3. The major legal AI platforms (LexisAI, TR Co-Counsel, and vLex) all declined to participate. I don't know why.
2. Human lawyers at the "top" US law firms finished well behind both types of AI.
3. The major legal AI platforms (LexisAI, TR Co-Counsel, and vLex) all declined to participate. I don't know why.
Nikki Shaver (@nikjet.bsky.social)'s summary of the report at @legaltechnologyhub.com is important: Nobody is saying that "AI is better than lawyers" in general terms. But there are still three significant takeaways from this report that really stand out for me:
October 17, 2025 at 3:23 PM
Nikki Shaver (@nikjet.bsky.social)'s summary of the report at @legaltechnologyhub.com is important: Nobody is saying that "AI is better than lawyers" in general terms. But there are still three significant takeaways from this report that really stand out for me: