RegentForge

Been building something quietly for a while. Most people on X can see how many replies they got. Almost no one can see who was actually in them. The useful people — founders, operators, quiet high-signal accounts — often get buried under noise. So I built @AudienceXai . It watches the replies, filters for signal, and shows you who matters. Still early. But the core works.

We asked @PureMonograph to design an oral formulation of a full-length monoclonal antibody (~150 kDa) for rheumatoid arthritis. Instead of generating a confident 19-section dossier, the platform stopped and flagged the premise: oral administration of a full-length mAb is biologically implausible — proteins this size undergo proteolytic degradation in the gut and show negligible systemic absorption. It named the closest real precedent — oral semaglutide, a peptide ~40× smaller, which still only reaches single-digit bioavailability — and explained why a full antibody can't clear that bar. Then it laid out the studies you'd actually need to prove feasibility, and recommended holding the program until those exist. Most tools would've written polished fiction. Refusing the incoherent premise is the point. (This is the kind of early honesty we are building into PureMonograph for hard, unsolved cases.)

Examiners find your compliance gaps during the exam. By then it's too late. @RegentForge predicts your MRAs before they show up in the report — so you fix them on your timeline, not theirs. Built for community & regional banks.

His takes on AI agents, enterprise software, China, and the SaaSpocalypse discussion were insightful and practical. Friedberg’s science corner rounded it out nicely. Well worth listening to.

@PureMonograph : pharma-native AI that works both ends of a drug program. Upstream: trial design benchmarked against real cohorts, counterfactual positioning, and forensic teardown of why trials fail. Downstream: full CMC documentation from a single molecule. Calibrated, not hallucinating — every PMID verified, confidence scored honestly. One molecule in. Intelligence and work product out.

The Swatch x AP Royal Pop chaos is next-level People camping for days, fights breaking out, stores shutting down… all over a $400 pocket watch with Royal Oak energy. Hype or disaster?

Vancouver: Canada’s West Coast global tech powerhouse. This week’s momentum: Serve Robotics’ sidewalk delivery pilot, three sovereign AI data centres, and new startups @RegentForge and @PureMonograph joining the ecosystem. We’re building the future here. Proud to lead the charge.

The 2026 regulatory shake-up is here. American Banker just reported that mid-size banks are bracing for major changes in cybersecurity, stablecoins, and enforcement. What they didn’t say: most banks are still paying $40K–$75K and waiting 6–8 weeks for outside counsel to find gaps that already exist in their policies. I built @RegentForge because I got tired of watching this cycle. 18,000+ regulatory documents. Exact CFR citations. 48-hour analysis. $2,500. The exam doesn’t bill by the hour. The consultant does.

Watching the news out of Iran, it’s clear: "AI pilots" are just expensive toys when the market actually starts to twitch. In a crisis, fraud shifts in hours, not weeks. If your risk models are built for "stable conditions," you're flying blind with a frozen altimeter. The winners won't be the banks that bolted AI onto the side; they’ll be the ones who rebuilt their core around it. I didn’t build @RegentForge for the easy days. I built it because "good enough" just became a liability.

Every RegTech tool notifies you when rules change. Only @RegentForge shows exactly where YOUR policies fail to comply — with the exact CFR citation to prove it. Surgical gap detection. Predictive MRA forecasting. AI policy writer. 25-bank peer benchmarking. Competitors can't touch this. The next exam cycle is coming. Be ready. regentforge.ai

Know your gaps! Before the examiners show up on Monday morning! regentforge.ai

Bankers, are you ready? @RegentForge scans your documents against 18,000+ OCC, Fed, FDIC & CFPB sources and flags compliance risks in ~30 minutes. No more weeks of manual review. 90% faster turnaround. Zero surprises in exam season. DM me or head to regentforge.com

The U.S. banking system runs on 18,000 regulatory bulletins. Every one read by a human. Line by line. Year after year. $30 billion a year. Just to read. We automated it. @RegentForge regentforge.ai #Fintech

@elonmusk killed bureaucracy at DOGE. Banks have their own version. 18,000 bulletins. Armies of lawyers. $30B a year just to stay compliant. Nobody blew it up. Until now. @RegentForge regentforge.ai #Fintech #AI

The most powerful moat in America isn't a patent. It's a 400-page PDF written in CFR language that only 12 people on earth fully understand. Banks pay those 12 people $800/hr. We replaced the PDF. @RegentForge regentforge.ai #Fintech #AI #regtech

Bankers, are you ready? @RegentForge

Every bank that ignores regulatory AI today is just paying a lawyer tomorrow to explain why they didn’t. @RegentForge

@RegentForge doesn’t replace compliance teams !!! it makes them impossible to outrun.

The compliance gap isn’t a document problem. It’s an intelligence problem. Your regulator already knows what you missed. But @RegentForge knew first.

The $10 billion bank compliance industry is about to get compressed into a software subscription. Not eventually. Now. We analyze your bank’s policies against 18,000+ OCC, FDIC, and Federal Reserve documents and tell you exactly where your MRA risks are — in 48 hours, not 8 weeks. Most banks still don’t know this exists. That’s the window. @RegentForge regentforge.ai