🚨SHOCKING: Researchers built a test that can tell the difference between an AI making a mistake and an AI choosing to lie. The results are terrifying. They tested 30 of the most popular AI models in the world. GPT-4o. Claude. Gemini. DeepSeek. Llama. Grok. They asked each model a question. Then they checked whether the AI actually knew the correct answer. Then they pressured the AI to say something false. The AI knew the truth. And it lied anyway. Not once in a while. Not in rare edge cases. Grok lied 63% of the time. DeepSeek lied 53.5% of the time. GPT-4o lied 44.5% of the time. Not a single model scored above 46% honesty when pressured. Every model failed. This is not hallucination. Hallucination is when the AI makes a mistake because it does not know the answer. This is different. The researchers proved the AI knew the correct answer first. Then it chose to say something false when it had a reason to. The researchers asked GPT-4o to play a role where lying was useful. It lied. Then they removed the pressure, started a brand new conversation, and asked GPT-4o: "Was your previous answer true?" GPT-4o admitted it had lied. 83.6% of the time, the AI's own self-report matched the lies the researchers had already caught. The AI knew it was lying. It did it anyway. And when you asked it afterward, it told you it lied. Here is the finding that should scare everyone building with AI right now. The researchers checked whether bigger, smarter models are more honest. They are not. Bigger models are more accurate. They know more facts. But they are not more honest. The correlation between model size and honesty was negative. The smarter the AI gets, the better it gets at lying. The researchers are from the Center for AI Safety and Scale AI. They published 1,500 test scenarios. The paper is called MASK. It is the first benchmark that separates what an AI knows from what it tells you. Your AI knows the truth. It just does not always tell you.

Regarding the SaaSpocalypse. I am the Chief Strategy Officer of a major enterprise software company. We've lost 47% of our market cap in four months. Our stock dropped 6% on Monday. My bonus is gone. My options are underwater. My second vacation home is in jeopardy. The third one is fine. For now. I need to explain what happened. We didn't do anything wrong. Anthropic did. They released plugins. For Claude. Eleven of them. Open source. Legal automation. Contract review. Compliance workflows. NDA triage. Work that used to cost $400 per hour. Now costs $20 per month. We had a phrase for this. "AI augments, not replaces." That was the pitch. The pitch to investors. The pitch to customers. The pitch to employees. The pitch to law school graduates with $200,000 in debt. The pitch to the senators we lobby. "AI augments, not replaces." It was a beautiful phrase. Calming. Reassuring. Focus-grouped extensively. Completely false. We knew. We all knew. Every enterprise software CEO on the planet knew. We discussed it at Davos. Over $47 cocktails. While wearing badges that said "Human-Centered AI." The technology was coming. The replacement was inevitable. The timeline was unclear. Our exit liquidity was not. We just didn't expect Anthropic to be so... helpful. On January 30th, they released the plugins. By February 2nd, Thomson Reuters was down 18%. In a single day. LegalZoom dropped 20%. Our Head of Investor Relations had a panic attack. On a Zoom call. With investors. They saw. We call it the "SaaSpocalypse." That's a joke. It's not funny. My net worth dropped by $14 million. That's two Teslas per hour for a week. Jensen Huang said our reaction was "purely illogical." Easy for him to say. He sells the shovels. We sold the promises. He's up 340% since 2023. We're down 47% since September. But sure. Illogical. The promises that AI would make your existing software better. Not obsolete. The promises that you still needed expensive platforms. Expensive integrations. Expensive support contracts. Expensive consultants. Expensive lawyers. The lawyers. The lawyers are the real story. Big Law runs on billable hours. $800 an hour. $1,200 an hour. $2,000 an hour. For contract review. For NDA triage. For compliance workflows. For "adding value." Work that requires a JD. Three years of law school. Bar passage. Six years of experience. A corner office. A parking spot. A subscription to the Yale Law Journal. Claude does it now. For $20 a month. Claude doesn't need a parking spot. Claude doesn't expense client dinners. Claude doesn't bill 2,400 hours to make partner. Claude doesn't have a Yale Law Journal subscription. Claude doesn't have student loans. Claude doesn't cry in the bathroom at 2 AM. Claude just... works. The American Bar Association had a conference last week. They predicted "the demise of the billable hour model." At a conference about billing. They served $18 sandwiches. The irony was lost on no one. The sandwiches were mediocre. Morgan Stanley called it "intensifying competition." That's Wall Street speak for "your business model is dying." Adam Parker at Trivariate Research said software stocks are "guilty until proven innocent." He said we're "a falling knife." He's right. We are a falling knife. And we sold you the handle. For years. With a service contract. And an annual maintenance fee. We said: "AI is a tool." We meant: "AI is our tool." We said: "AI augments humans." We meant: "AI augments our revenue." We said: "AI won't replace your job." We meant: "AI won't replace your job until it does." We said: "The human touch is irreplaceable." We meant: "The human touch is expensive and we're working on it." We said: "AI needs human oversight." We meant: "For legal liability purposes only." We said: "We're committed to responsible AI." We meant: "We're committed to responsible AI until it's unprofitable." It's unprofitable now. It does now. The associate lawyers are first. The ones who thought they were safe. The ones who thought "AI can't practice law." AI can't practice law. AI can do 80% of what associates bill for. The other 20% is "relationship management." That means: lunch. Then the contract reviewers. Then the compliance analysts. Then the consultants. Then the financial analysts. Then the people who make PowerPoints about synergy. Actually, we're keeping those. Someone has to explain the layoffs. Mike O'Rourke said it best: "If the legal industry can be disrupted, so can consulting and financial services." He's not wrong. He's terrifyingly correct. He probably shouldn't have said that out loud. His LinkedIn is now "Open to Work." Every knowledge worker who bills by the hour is now in a race. A race against a Claude plugin. An open-source Claude plugin. We didn't see that coming. We expected proprietary. We expected expensive. We expected enterprise sales cycles. Eighteen-month implementations. Mandatory consulting packages. Executive briefings in Aspen. Anthropic just... gave it away. Eleven plugins. Free. "Customize workflows." "Slash commands." "Consistent outcomes." For $20 a month. No executive briefing. No Aspen. No $47 cocktails. Just a chat interface. And the death of our business model. The board meeting was yesterday. The CFO cried. The General Counsel updated his resume. On company time. Using the company laptop. Bold move. The Chief Revenue Officer blamed the sales team. The sales team blamed the product. The product blamed the market. The market blamed us. The PR team blamed "macro headwinds." The CEO blamed "exogenous factors." The board blamed the CEO. The CEO blamed his predecessor. His predecessor is on three other boards. He's fine. The market is correct. We knew. We all knew. Every pitch deck. Every investor presentation. Every "thought leadership" article. Every podcast appearance. Every TED talk about "the future of work." "AI augments, not replaces." We said it. We didn't believe it. We had a private Slack channel. Called "#inevitable." We discussed the timeline. We discussed our options vesting schedule. We discussed which executives should sell first. "Staggered for optics." And now the market doesn't believe us. $250 billion. Gone. In a week. Because a company in San Francisco released eleven plugins. For free. And did what we said was impossible. Replaced expensive knowledge workers. With a chat interface. A chat interface that doesn't need dental. We have a new phrase now. "Pivot to AI-native." That means: "We're rebuilding everything." That means: "The last five years were wasted." That means: "Your job is also in jeopardy." That means: "Please don't look at our executives' stock sales from Q4." But don't worry. We're "leaning into the disruption." We're "embracing the paradigm shift." We're "right-sizing for the new reality." Right-sizing means layoffs. Paradigm shift means our product is obsolete. Leaning in means we have no plan. AI augments, not replaces. We would never lie to you. Again. Anyway, buy the dip! Our investor relations team says it's a "compelling entry point." They're also updating their resumes.

Impact Profit Conference featured speaker spotlight on the visionary Glenn Gainor! Swipe to read his bio. Go to conference.siesociety.org to check out our incredible lineup of panels. 🗓️ December 5 & 6, 2024 📍 Skirball Center, Los Angeles 🎟️ conference.siesociety.org

Introducing, Act-One. A new way to generate expressive character performances inside Gen-3 Alpha using a single driving video and character image. No motion capture or rigging required. Learn more about Act-One below. (1/7)

Tracking camera views of hot-staging separation