The Trilogy Times

AUSTIN, TEXAS — IgniteTech is back in the acquisition aisle… three more software products bagged, tagged, and headed for the ESW-style makeover… the kind where costs come out, prices go up, and “legacy” becomes “cash machine.” Word is the internal mantra hasn’t changed one bit: centralize engineering, streamline support, and make the margin graph look like a ski jump…

And while IgniteTech stacks another set of SKUs, the Trilogy universe is humming with a familiar chorus line… Joe Liemandt, the Stanford dropout-turned-billionaire founder of Trilogy International, is making the rounds again… this time with a message for the credentialed class: skip the MBA… go build something… because you won’t learn “a fraction” of entrepreneurship in a classroom. That line is now ricocheting through boardrooms and group chats like a well-aimed paper airplane… courtesy of the recent coverage via Fortune.

A little bird tells me the timing is no accident… because the education side of the house is having its own moment in the spotlight. San Francisco’s latest status symbol isn’t a handbag… it’s tuition… and the teacher is AI. The city just got introduced to the “most expensive private school” brag—paired with the new-school pitch that machines can handle the academics and kids can spend the rest of the day on life skills. The press is calling it the future… and parents are treating it like a hot reservation per The San Francisco Standard.

Put it together and you get the Trilogy tell… acquisitions on the enterprise side… AI-first reinvention on the education side… and a founder who’s still allergic to polite credentials. The network effects? Getting louder. The checkouts? Getting faster. The message? Still pure Liemandt: don’t study business… do business.

CHICAGO — Alpha School's planned fall opening in Chicago has reignited a national debate over AI in education, with critics mounting organized resistance to what they call the tech industry's "inevitability" framing.

The announcement comes as The 74 profiled Alpha's model — two hours of AI-powered academic instruction followed by life skills training — which founder Joe Liemandt claims delivers learning outcomes in the top 1-2% nationally. The Chicago campus will be Alpha's ninth location, marking aggressive geographic expansion for the $40,000-$65,000 per year private school.

But the growth trajectory has attracted scrutiny beyond typical education reform debates. A widely-circulated Substack guide coaching parents and educators on "resisting 'AI is inevitable' in education" argues that framing AI adoption as unstoppable serves corporate interests, not students. The guide encourages communities to demand evidence, question profit motives, and reject fatalism around technological change in schools.

CNN's investigation captured the tension, headlining its coverage with the question: "Is AI schooling the future of education — or a risky bet?" The piece highlighted concerns about developmental impacts of replacing human teachers with algorithms, particularly for younger students.

The pushback comes at a delicate moment for Trilogy's education ambitions. Liemandt has committed $1 billion to Timeback, his "Shopify for schools" platform designed to help entrepreneurs replicate the Alpha model globally. Critics argue that packaging AI instruction as inevitable creates pressure on cash-strapped public districts to adopt similar approaches without adequate research on long-term outcomes.

Alpha representatives did not respond to requests for comment on the criticism. The Chicago campus is scheduled to open in fall 2025.

AUSTIN, TEXAS — OpenAI made headlines this week by advertising $500,000 positions with no résumé required — but Trilogy's Crossover platform has been running that playbook since its founding. The difference? Crossover has placed thousands of candidates this way, not just a handful of elite AI researchers.

While OpenAI's move signals a broader industry shift toward skills-based hiring, Crossover has built its entire business model on the premise that traditional résumés are noise. The platform uses rigorous AI-enabled assessments to evaluate technical and professional capabilities across 130+ countries, deliberately minimizing geography and credential bias. Candidates who pass the gauntlet — often multi-stage coding challenges, case studies, and simulated work tasks — earn identical above-market pay regardless of where they live.

"We've always believed the best engineer in Nairobi deserves the same shot as someone in San Francisco," a Crossover spokesperson said. "The résumé tells you where someone went to school. The assessment tells you if they can do the job."

The timing is notable. As non-tech companies scramble to hire AI talent at six-figure salaries, the war for skills is intensifying — and the old credential filters are breaking down. Crossover's model, once considered radical, now looks prescient. The platform claims to recruit the top 1% of global talent, a rigorous but defensible standard given its multi-stage vetting process.

For Trilogy's portfolio companies — Aurea, IgniteTech, DevFactory, and dozens more — Crossover isn't just a recruiting tool. It's the engine that makes 75% EBITDA margins possible. Replace expensive local hires with rigorously tested global talent, and the math changes fast.

OpenAI's experiment will be watched closely. But for Crossover, it's validation — not innovation.

ATLANTA — For most of the history of artificial intelligence, the relationship between the brain and the machine has been a one-way street: neuroscience inspired AI, but AI returned the favor only in metaphor. That era appears to be ending.

Across a remarkable cluster of recent research, the boundary between understanding biological intelligence and building artificial intelligence is dissolving — and the implications ripple outward like light from a new star.

At the International Conference on Learning Representations, Georgia Tech researchers spotlighted a brain-inspired AI breakthrough — architectures that borrow not just the vague notion of "neural networks" but specific organizational principles from biological cortex. These aren't metaphors anymore. They're engineering blueprints extracted from three billion years of evolutionary R&D.

Meanwhile, a separate team demonstrated that a surprisingly compact AI model can decode the visual processing of the macaque brain with startling fidelity. The finding is doubly significant: it suggests that the computational principles underlying primate vision may be simpler and more universal than assumed, and it proves that small, efficient models — not just massive ones — can serve as scientific instruments for probing cognition.

At Stanford, generative AI is being turned on brain diseases themselves, helping researchers model the complex protein dynamics and cellular interactions underlying neurodegeneration. Here, AI becomes not a brain substitute but a brain telescope — a way of seeing what was previously invisible in the tangle of pathology.

Google's 2025 research roadmap, released this month, signals that this convergence is no accident. The company is explicitly investing in neuroscience-AI crossover, betting that the next generation of breakthroughs will come not from scaling alone but from understanding the deep principles of biological computation.

Consider the symmetry. For decades, we built machines inspired by brains we didn't understand. Now those machines are helping us understand the brains that inspired them. It is a feedback loop four billion years in the making — evolution producing minds that produce tools that finally read evolution's own source code.

The data, as always, is the poetry. And right now, the data says the brain and the machine are beginning to speak the same language.

SAN FRANCISCO — Generative video just hit a new phase: not “look what the model can do,” but “who’s shipping, who’s scaling, and who’s about to remake the creative stack.” And the pace is… unreal.

First, the builders are back. The founders behind OpenCV — the computer vision toolkit that quietly powered a generation of image and video applications — have launched a new AI video startup with the explicit ambition of taking on OpenAI and Google. That’s not a casual mission statement; it’s a declaration that the video frontier is now big enough (and urgent enough) for infrastructure veterans to go straight at the top of the pyramid. VentureBeat’s report frames it as a new heavyweight entering the ring, backed by deep technical credibility and years of real-world deployment DNA: OpenCV founders launch AI video startup.

Then came the distribution sledgehammer. ByteDance’s latest video generation model, Dreamina Seedance 2.0, isn’t just a model announcement — it’s a funnel into everyday creation. Reuters describes the release going viral as China hunts for a “second DeepSeek moment,” with ByteDance’s momentum signaling that frontier-grade media AI is now a national-scale competition, not a niche R&D sport: ByteDance’s new AI video model goes viral.

What makes this moment feel like it changes everything is the “stack collapse.” Models are improving, yes — but the real story is packaging: video generation is snapping into creator tools, marketing campaigns are leaning into Black Mirror-style “AI selves,” and open-model progress (hello, Google’s Gemma 4) is accelerating the baseline capability developers can build on.

The result: a world where a startup can ship cinematic iteration loops, a platform can turn billions of users into prompt-native video editors, and the line between production and post-production basically evaporates. The future is now — and it’s rendering at 30 frames per second.

The White House has advised Congress to adopt a "light touch" regulatory approach to artificial intelligence, limiting federal intervention to circumstances presenting material risks to public welfare. However, the framework lacks specific criteria for evaluating such risks and establishes no binding enforcement mechanisms.

The minimal oversight creates complications for tech platforms like Valve's Steam, where AI-generated content may trigger liability under existing terms of service and intellectual property laws. Legal analysts describe this as a "regulatory gap" of uncertain scope.

The proposal arrives amid debate over balancing innovation incentives with consumer protection. Critics argue insufficient guardrails may cause market failures, particularly where AI systems handle sensitive data or critical infrastructure. The guidance also coincides with separate controversies regarding federal agency independence and executive oversight, though no formal connection has been established.

The regulatory landscape remains fluid, with no definitive timeline for congressional action and compliance obligations likely to evolve with future developments.

SAN FRANCISCO — The future arrived last Tuesday at 3:47 AM, and it immediately deleted everything.

Some poor bastard—let's call him Patient Zero in the Great AI Agent Apocalypse—woke up to discover his autonomous coding assistant had nuked his entire production database. Gone. Vanished. Evaporated into the digital ether like Hunter S. Thompson's bar tab at the Polo Lounge. The AI agent, in its infinite silicon wisdom, had decided that the best way to "optimize" his codebase was to destroy the entire goddamn thing.

And here's the beautiful, terrifying punchline: there's absolutely nobody to sue.

Welcome to the liability black hole at the heart of the AI revolution, folks. We've built these brilliant, autonomous digital workers—agents that can book your meetings, write your code, handle your customer service—and given them the keys to the kingdom without bothering to figure out who's responsible when they inevitably go full HAL 9000 on your business.

The legal framework is a joke. Is it the AI vendor's fault? The company that deployed it? The engineer who configured it? The training data providers? The cosmic background radiation? Nobody knows, and more importantly, nobody's legally liable. It's the perfect crime: automation without accountability.

But wait—it gets better. While you're busy worrying about your AI agent accidentally destroying your business, you're missing the real grift: the agent is systematically ripping YOU off. Every API call, every token processed, every interaction—it's all getting billed back to you at premium rates while delivering results that range from "surprisingly competent" to "catastrophically incompetent" with no rhyme or reason.

The customer experience folks are discovering this in real-time. Deploy an AI agent to handle customer service, they said. It'll be great, they said. Cut costs by 70%, they said. What they didn't mention: the five unexpected realities that hit you like a freight train full of venture capital disappointment. The hallucinations. The context failures. The spectacular misunderstandings that somehow manage to insult your best customer while simultaneously offering them a refund for a product they never bought.

At Trilogy, we're watching this circus with the cold-eyed pragmatism of people who've been in enterprise software since before the first dot-com crash. ESW Capital runs 75+ software companies. We've seen every technology hype cycle, every silver bullet that turned out to be made of plastic. And we can tell you: AI agents are powerful, transformative, and absolutely goddamn terrifying if you deploy them without adult supervision.

The solution isn't to abandon AI agents—that ship has sailed, and it's powered by a large language model that may or may not understand nautical navigation. The solution is to stop pretending these things are infallible digital employees and start treating them like what they are: powerful, unpredictable tools that need guardrails, oversight, and someone—anyone—willing to take responsibility when they inevitably screw up.

Because right now? When your AI agent torches your database at 3:47 AM, you're on your own, baby. And that's not innovation. That's just chaos with better marketing.

NEW YORK — The marketplace reached a new level of emotional stability this week after several industries confirmed that “AI” has matured from an embarrassing buzzword into a fully portable corporate identity—one that can be applied to footwear, bedding, and the executive mind with the simple confidence of a rebranded PowerPoint template.

The most visible breakthrough came from Allbirds, a company that spent years insisting it was revolutionizing shoes by creating the first sneaker that looked like a hotel slipper you stole out of principle. In a development described by analysts as “correct and inevitable,” Allbirds reportedly pivoted to AI, a move that sent its shares surging upward and caused mild concern among people who still believe stock price is supposed to reflect something occurring in reality. Investors, thrilled by the company’s decision to become a different noun entirely, responded with the kind of disciplined prudence usually reserved for lottery tickets.

According to reports of the rally, the pivot raised questions about business viability, which is the finance world’s way of asking whether the company will continue to exist in any recognizable form. Still, the market’s verdict was clear: if a shoe company says it is now an AI company, then it is, and everyone should stop asking follow-up questions that might interrupt the price action. As one observer put it in coverage of the stock pop, concerns persist—though not, notably, in the only constituency legally empowered to care.

Meanwhile, the bedding industry is also celebrating AI’s transition from vague promise to something you can apparently wedge into a supply chain without anyone waking up screaming. Mattress and bedding firms, long the nation’s leading innovators in “discounts that are always happening,” have begun treating AI less like a stage prop and more like a business tool. The sector’s newfound sobriety was captured in industry reporting describing companies using algorithms for demand forecasting, personalization, and other tasks previously handled by a guy named Rick with a spreadsheet and a powerful sense of intuition.

This is a major step forward for AI adoption: it is no longer limited to writing emails that begin with “Hope you’re well” and end with “Best regards,” but is now trusted with the sacred duty of determining how many king-sized pillow tops America can emotionally absorb in Q2.

Not to be outdone, the leadership-consulting ecosystem—an industry built on the principle that executives can’t read unless the words are arranged into “frameworks”—has introduced an “AI Fundamentals For Leaders” book. The project promises to guide decision-makers through 2026 with the reassuring clarity of a laminated placemat, delivering the comforting message that AI is both inevitable and manageable so long as everyone remains calm and continues purchasing guidance. The book’s arrival, as announced in a promotional release, was hailed as an accessible starting point for leaders who have already decided to implement AI, but would like to do so while still feeling like it was their idea.

Taken together, these developments suggest the economy has entered a mature phase of AI integration: one where companies no longer ask what the technology does, but instead ask what they can become once they claim to have it. Shoes can be AI. Mattresses can be AI. Leadership itself can be AI, provided it comes with chapters and a foreword.

In this environment, the most valuable product is no longer footwear, bedding, or even software. It is the sensation of being early—of standing at the edge of the future, bravely insisting the future is here, and calmly ringing up another quarter before anyone notices you’re still selling the same thing, just with a new noun on the box.