Raad @R44D

founder/ceo @lawtrades. freelance legal marketplace. building something new🦞 raad.substack.com nyc Joined April 2009

Tweets

9K
Followers

17K
Following

1K
Likes

14K

Raad @R44D

10 hours ago

@gabepereyra @charle_marle @harvey @Mercor @SnorkelAI if you need more help, attorneys on @Lawtrades doing eval work with live client context. fastest growing thing we do

1 0 2 116 0

View Details

Brickell Finance Guy @IronChestCap

4 weeks ago

me with the SpaceX cafeteria lady after she becomes a billionaire from the IPO

438 3K 39K 4.7M 2K

View Details

Raad @R44D

a month ago

@gabepereyra why don’t you guys just start a law firm on the side?

1 0 0 364 0

View Details

“ok this startup is cool but …” 1980: … what if IBM builds this? 1995 … what if Microsoft builds this? 2010 … what if Google builds this? Today … what if builds this? reality is, if founders listened to the “what if” pessimists we’d never have any startups or new products. That’s why they’re building and the pundits aren’t My observation: When these huge waves happen, these new markets are so damn big there will be tens of thousands of new viable companies, hundreds of unicorns, and a few iconic companies that become generational. The big cos play a role but can never compete with the glorious open market known as capitalism So for all the “what if” people - sit down, log off X for a bit, and let the founders do their thing. And let’s cheer them on when they do

142 125 1K 85K 291

View Details

Raad @R44D

3 months ago

@NotionHQ @ivanhzhao @claudeai @AnthropicAI super clunky experience

0 0 0 346 0

View Details

Raad @R44D

3 months ago

durable task flows = your agent survives a crash and picks back up from where it stopped. no lost work, no restart from scratch. exec approvals = you stop watching every command run and only get pulled in when something actually needs a human call. those two changes alone make agents feel like infra instead of demos.

OpenClaw🦞 @openclaw

3 months ago

OpenClaw 2026.4.2 🦞 🔄 Durable Task Flow orchestration 🔓 Better native exec defaults + approvals 🤖 Copilot + Kimi + provider hardening 🔌 Tighter plugin activation boundaries 🛡️ Hardened provider transport + routing Less bloat. More lobster. github.com/openclaw/openc…

197 185 2K 235K 324

4 0 4 383 1

View Details

Raad @R44D

3 months ago

@balajis iran has no incentive to stop

0 0 1 68 0

View Details

Raad @R44D

3 months ago

@signulll teaches human system optimization. nothing about when your ai agents start optimizing against you overnight. no kill-switch module in any curriculum.

0 0 0 68 0

View Details

Raad @R44D

3 months ago

@danmartell if your friends are talking about claude cowork instead of open claw… you need new friends.

0 0 0 57 0

View Details

Raad @R44D

3 months ago

@AlexFinn the gap isn't mythos vs sonnet. it's the people who've implemented ai vs the people who haven't even started. you're right to be nervous. mythos just widens something that's already open.

0 0 2 422 0

View Details

Raad @R44D

3 months ago

@karpathy your llm council idea helped me tremendously with this concept

0 0 1 206 1

View Details

Raad @R44D

3 months ago

@roshanchokshi been experimenting with paperclip. goal-scoped agents that report up instead of running indefinitely. closer to what actually works.

1 0 1 27 0

View Details

Raad @R44D

3 months ago

nobody talks about agent drift on long runs. models degrade, context dilutes, outputs silently veer. 24/7 setups need checkpoints, not just autonomy. that's where the real value is.

Martin Shkreli @MartinShkreli

3 months ago

what is the best tooling for 24-7 inference/agent-driven research? im trying factory but it stops and asks me questions even though i have 'auto' mode on. tbh i think this is an even bigger killer app than LLM chatbots. who else is out there doing it?

150 21 582 128K 571

3 0 3 261 0

View Details

Raad @R44D

3 months ago

@pmarca the unspoken part: agents in prod don't replace labor, they concentrate it. hiring surge is in orchestration. humans becoming the conductor not the musician. every agent deployed 3-5x the oversight/iteration/monitoring headcount. the jobs aren't gone, they're harder to build.

0 0 0 227 0

View Details

Raad @R44D

3 months ago

the problem nobody building multi-agent systems will admit: the agents don't fail. the coordination does. one agent hallucinates. another acts on stale data. the third makes an irreversible call. you don't catch it until it's already downstream. "autonomous agents" is a marketing term. practitioners call it ghost management.

0 0 1 138 0

View Details

Raad @R44D

3 months ago

@DietCoke_Esq 🎯

0 0 0 7 0

View Details

Raad @R44D

3 months ago

the best legal departments figured out something most haven't. the lawyer's job isn't just to protect the company. it's to make every internal employee obsessed with working with legal again.

2 0 3 294 0

View Details

Raad @R44D

3 months ago

lawyers don't want another tool to log into. the contract review feature getting the most traction: cc an email address. get a redline back. that's it. no new app. no new workflow. just email.

0 0 1 261 1

View Details

Raad @R44D

3 months ago

pasting a prd into an ai and saying "build this" doesn't work. the job is breaking it into phases small enough that an agent can't fail.

0 0 1 209 0

View Details

Raad @R44D

4 months ago

every founder i know has a version of the same habit: keep a notes app, a notion page, a slack thread, and a voice memo. all for the same thought. we weren't disorganized. we were waiting for something that could hold context the way we do. that's what an actual AI teammate does.