Adan (@ordonez_adan) 's Twitter Profile
Adan

@ordonez_adan

Real Estate|Tech|Law @uchicagolaw

ID: 2843928804

linkhttp://lawbandit.com calendar_today07-10-2014 17:41:15

1,1K Tweet

632 Followers

839 Following

Nikunj Kothari (@nikunj) 's Twitter Profile Photo

Every time I see a tweet saying “I can vibe code this in a weekend” - I think of the slack notification system.. It takes time, persistence and effort to get the details right. Sure, a lot of simple workflows will get vibe coded away. And maybe you can put this in Claude Code

Every time I see a tweet saying “I can vibe code this in a weekend” - I think of the slack notification system..

It takes time, persistence and effort to get the details right. 

Sure, a lot of simple workflows will get vibe coded away. And maybe you can put this in Claude Code
Adan (@ordonez_adan) 's Twitter Profile Photo

One of the most valuable skills in the AI era is knowing when the output is wrong. AI gives everyone access to better starting points. Both good and bad lawyers benefit from AI. But the gap between them doesn't close, because the good lawyer still has better judgment about the

Adan (@ordonez_adan) 's Twitter Profile Photo

Good video. I have a feeling that skills will be the new apps. Kind of like we have the Apple App Store. People using the regular AI models with connection to MCP/skills to do things with it. Already happening but I think this will take over.

SMB Attorney (@smb_attorney) 's Twitter Profile Photo

What if… and hear me out now… instead of paying $5K per license for Claude with a dopey wrapper, you just pay $200 and use Claude?

What if… and hear me out now… instead of paying $5K per license for Claude with a dopey wrapper, you just pay $200 and use Claude?
Adan (@ordonez_adan) 's Twitter Profile Photo

Can anyone fact check this? I wonder if people actually throw that much harder now or if the way velocity is tracked skews actual velo that much

Adan (@ordonez_adan) 's Twitter Profile Photo

Many on LinkedIn and X are posting sloppy AI-generated content that gets a lot of engagement but is inaccurate or missing important context - clickbait essentially. People often send me headline posts about which model is best or what works and what doesn't, and clearly the

Adan (@ordonez_adan) 's Twitter Profile Photo

Some of my professors sign off on their emails with their initials. I think it looks cool. At what point is it acceptable to sign off emails with your initials? -ACO

Adan (@ordonez_adan) 's Twitter Profile Photo

After being a non-lawyer my whole life then going to law school I can confidently say this is an awful idea. Before law school I did not understand anything about legal work. For example I guarantee most don't know what "boilerplate" even means. I didn't. And I wouldn't

Adan (@ordonez_adan) 's Twitter Profile Photo

As a law student, I can already see we're approaching the law differently than current associates or partners. Even between my 1L and 2L year, there's been a big shift. A lot of my peers use AI first for studying, drafting, and researching. This makes me think that there's no

As a law student, I can already see we're approaching the law differently than current associates or partners.

Even between my 1L and 2L year, there's been a big shift. A lot of my peers use AI first for studying, drafting, and researching. This makes me think that there's no
Adan (@ordonez_adan) 's Twitter Profile Photo

The problem with AI errors is that one small mistake can snowball. If you don’t catch the error early, the model will keep building on it and keep generating more wrong information based on that first error. For example, if an LLM hallucinates a case and you keep asking

Adan (@ordonez_adan) 's Twitter Profile Photo

The best way I've found to create a skill in Claude is to get Claude to do the task first. Once the output is good, tell Claude to package the whole thing into a skill. I've done this across legal research and drafting workflows. Lots of trial and error - the skill is the end

Adan (@ordonez_adan) 's Twitter Profile Photo

Firms/Laywers adopting AI have to accept that your work might be worse at first while you figure it out. AI is not deterministic: Similar inputs get a variety of outputs. Getting the prompting and the workflow right takes trial and error, which equates to inefficient work that

Adan (@ordonez_adan) 's Twitter Profile Photo

Hiring a lawyer for a contract is insurance. Could things still go wrong with a lawyer involved? Yes. But you have a much lower chance. The alternative is drafting something on ChatGPT or Claude and sending it to the other party, hoping it holds. If it goes wrong, it could go

Adan (@ordonez_adan) 's Twitter Profile Photo

AI can pull relevant law faster than any lawyer. So why would anyone still pay $800 an hour for legal advice? The answer is that AI tells you what the law is and may give you suggestions on what to do with it, but a good lawyer can verify the law and apply judgment to it. Heck,