John Hellerman

@nicksortor @elonmusk They’re calling it a loss for Elon, but he’s probably made back his lawyer fees already today, and I would say Sam Altman‘s reputation is in the crapper.

🚨 JUST IN: A jury has TOSSED Elon Musk’s lawsuit OpenAI and Sam Altman, after Altman allegedly STOLE the non-profit AI company Elon bankrolled, and turned it into a for-profit venture The jury did NOT say @ElonMusk’s claim was invalid, but said it was filed too late. Insane. h/t @iam_smx

@atlblog @Cmbg3Law This is an important discussion. Without the wisdom of “doing the work” now it’ll be difficult to know in the future what to even ask for in a prompt.

AI speeds things up, but should it think for us? 🎧 Attorney Harshita Ganesh from @Cmbg3Law discusses cognitive offloading risks for junior lawyers and AI's impact. Watch: ow.ly/ZBHE50YO3l7

The smallest mistake...when I first started my career doing PR for lawyers I went to work for @JaffePR -- their circa 1996 website's home page started with this: One "oops" can erase a thousand "well done's"! And that was the ethos of the work -- you got it perfect bc if one thing wasn't, none of it could be trusted.

It was 1am during one of several all-nighters I pulled working on a complex cross-border finance deal. I was a first year associate at Debevoise and the senior associate on the deal strode into my office and dropped off some markups for me to process. She flipped to one of the 27 sets of signature page packets I had prepared and pointed emphatically at one of my mistakes, which she helpfully encircled using half a red Bic's worth of ink. In one of the signature blocks, I spelled the entity name as "ENTITY HOLDINGS, INC." when in fact it should have been "ENTITY HOLDING, INC." "Clients pay us what they pay us so they can sleep soundly knowing that we will execute with dogged perfection" she explained. The smallest mistake, however insignificant, was an indicia that the higher stakes elements of the deal may not be airtight either. And that would be fatal to a client's confidence in your work product. The craft was the craft, regardless of whether it was a critical covenant or an entity name on a signature page. 13 years later I still think about that night. It's easy to dismiss the tasks of a junior associate as valueless, menial grunt work, but the unglamorous work compounds in counterintuitive ways. Checking and rechecking section cross-references, tracing through a maze of nested defined terms, and yes--compiling signature pages--were often forcing functions to do more important things like *actually* reading a complex contract from top to tail. Over time, these thankless tasks built a rich latticework of mental models and a pattern-matching library that to this day I draw upon when facing new and unusual legal issues. Even when using AI (which I do every day), this decade-long scaffolding of knowledge I've cultivated helps me ask the right questions, prod with the right followups, and curate and assemble the best answers. I'm sure, just like generations of old men futilely yelling at clouds before me, I'll be proven wrong, but man, I'm worried (and a little sad) for the new generation of lawyers.

This is NOT an episode of Weeds.

I don;t think it’s any of our business who the culprit is. That’s an internal matter. Whether true or not, no believes a @sullcrom partner was the one responsible for the actual work and “back in your day” and now, whoever it was, we’d have to believe, will never touch another piece of work at the firm.

"In my day, an associate who submitted invented citations would be gone by lunchtime. No debate." We have yet to hear what has happened to the person directly responsible for irresponsible AI use. That would tell us a lot.

BigLaw at its best and worst in one filing. Sullivan & Cromwell owned the issue immediately, which is impressive and important. The scar, though, is real: when you charge what they do, clients assume zero mistakes. Peers will whisper, but if @sullcrom continues to be transparent about the new tools and protocols they’re implementing, they'll turn a PR hit into a trust win.

Elite law firm Sullivan & Cromwell admits to AI ‘hallucinations’ ft.trib.al/wrVB84e

Alex, you're spot on as always. While clients expect better diligence than a solo at $125/hr., the letter was genuine accountability: painful but classy. I believe that @sullcrom's transparency on fixes will earn back trust fast. No one wants an error but how you respond to a screw-up really does define you.

I'm sure it was incredibly painful to write this letter. Short term, you'll absolutely get roasted by the judge & by everyone. It's a PR nightmare. And yes, you may even get fired. But by owning responsibility for the mistake and taking the initiative to find *other* unsurfaced mistakes to mitigate them, you take one small step towards earning that trust back. In my opinion, how you respond when you mess up is a huge sign of whether or not you can be trusted in the future.

This will be uncomfortable for them for a while (esp. w/all the gossip and peer whispering) but it's @sullcrom and how they handle this now on their own behalf -- and so far so good (transparent, apologetic, etc.) -- will demonstrate how they handle similar troubles at $3K per hr.

Picked up by the MSM. Yikes.

POV: You just paid S&C, one of the three most expensive and high-powered law firms in the world, $3000 per hour to submit AI slop to the court on your behalf. No one is safe.

While peer whispering is inevitable and will be uncomfortable for a while, Sullivan & Cromwell's proactive internal review and swift apology letter from Dietderich was textbook accountability and exactly how a class firm should respond when things go wrong. That said, this is still a serious self-inflicted wound for a firm as prestigious (and expensive) as @sullcrom, and it will no-doubt cause serious internal reviews and new mitigation protocols. Ultimately, S&C's handling of the mess will say more about the firm than the mess itself.

Sullivan & Cromwell, a premier Wall Street law firm, apologized to a federal judge for submitting a court filing with inaccurate citations and other errors generated by artificial intelligence. reuters.com/legal/litigati…

A speech Trump should finally give on Iran from 2009!

6/6. If you need credible earned media but can't afford "to try" -- DM me about our results-based FAME program for professionals.

5/6. Are you seeing this shift in your 2026 budgets and planning yet? What’s your biggest challenge right now in prioritizing (or scaling) earned media? Getting buy-in to reallocate from owned/paid? Building deeper reporter relationships? Measuring real AI-era impact?

Good morning ☀️ Saw this eye-opening stat from Gartner’s latest “Top Predictions to Inform 2026 Comms Strategies” report: The mass adoption of AI and public LLMs (Grok, ChatGPT, Copilot, Gemini, etc.) is changing everything. As AI becomes the new primary search interface, it’s expected to drive a ~2x increase in PR and earned media budgets by 2027. Why? Because these “answer engines” overwhelmingly pull from — and cite — credible journalism and earned media. Gartner notes that more than 95% of the links they surface come from non-paid, authoritative journalistic outlets. In short: real PR placements in respected publications now carry far more weight than another blog post or LinkedIn article. Brands without substantial earned media risk becoming invisible in AI-powered results. 1/6

@StockMKTNewz Congrats to @PalmerLuckey We’re lucky to have someone of his resilience and vision leading our innovation efforts in military and war fighting.

ANDURIL IS NOW WORTH $60 BILLION Thrive Capital and Andreessen Horowitz are co-leading a multibillion-dollar investment round in Palmer Luckey founded defense startup Anduril Industries that doubles the company’s valuation UP TO $60B, including the investment - WSJ

It's Purim tonight and Esther may be history’s most underrated communications strategist. Timing. Coalition-building. Message discipline. Still relevant for leaders today. Here's my take on Esther the Influencer: hellermanllc.com/effective-comm…

If you ask your Q like this “Does allowing AI to run over Shabbat preserve my existential stillness, or does it extend my weekday agency into sacred time?” You get an answer like it shouldn’t be “your agent” — it shouldn’t be acting as you would on real time inputs, especially in a public way (eg, answering email, posting to X, etc.)

Not enough people understand this like 99% of the Talmud

Can a Jew let his agent run over Shabbat if the last prompt was Friday afternoon?