Max Armbruster

"One of the simplest, highest ROI thing to put an agent on? Your Contact Me form. No one should ways days to hear back from a 21 year old human to qualify them."

Credit where it's due: Accenture booked $5.9 billion in AI projects last year. Nearly double the year before. This joke only works if Accenture is a code shop (It's not). Accenture pulls in $70 billion a year. Half of that ($34.6B) is managed services, where they run entire business operations for other companies: payroll, HR, supply chains, security, compliance, all of it, across 120 countries. The consulting half is mostly strategy and organizational overhauls for 9,000 clients, not writing code. The irony: every time a company wants to deploy AI tools internally, they call Accenture. Someone has to figure out which workflows to automate, retrain thousands of people, restructure processes, migrate data, and bolt AI into enterprise systems built decades ago. That's exactly what Accenture sells. And it's working. Accenture's AI revenue tripled to $2.7 billion in fiscal 2025. They've run 11,000 AI projects across 1,300 clients since 2023, totaling $11.5 billion in AI bookings. Last quarter, they booked another $2.2 billion in AI work alone, with AI revenue up 120% year over year. Total quarterly bookings hit a record $20.9 billion, with 33 clients each signing $100M+ deals in a single quarter. They did cut about 11,000 employees who couldn't be retrained for AI. But they grew their AI workforce from 40,000 to 77,000 in two years and trained 550,000 employees on gen AI basics. Their CEO said AI is now so embedded in everything they do, they're going to stop reporting it as a separate number. The real risk to Accenture is that AI makes its own consultants so productive that clients stop paying for hours and start paying for outcomes. Revenue per employee sits at ~$89K/year, suggesting that much of their workforce performs the kind of process execution and staff augmentation that AI genuinely compresses. Their stock is down 33% from its 52-week high, even after beating earnings, because Wall Street sees the same math. A 24-year-old with Claude Code can ship an app over a weekend. Accenture charges Fortune 500 companies to rewire how they operate across 52 countries. Those are different jobs. AI tools are making Accenture's phone ring more right now, but the margin pressure from those same tools is the real story to watch.

Met an American founder at a tech conference in Amsterdam He showed me his website and I noticed there is no cookie banner “Ah yeah, we didn’t include it, it’s just annoying and no one cares anyway” My blood started to boil I excused myself and immediately called the Dutch Data Protection Authority They connected me to Europol At 4am the next morning, police raided his Airbnb and he called me crying from the station “What the hell bro, did you really call the cops on me for a cookie banner?” I felt bad for about 25 seconds but then I remembered: we are in Europe and rules are meant to be followed He is being extradited tomorrow back to the wild west called America Sometimes being a good founder means making the hard calls

So the AI race boils down to: Who can afford to pay the most for Chinese dudes and who can buy more GPUs from another Chinese dude, and can the Chinese dudes in China keep up?

Good morning, India! 🇮🇳❤️ While the general mindset around building a business is to move fast and break things, we asked, how do you stay the course and build one for the long term? At @Zoho, we chose an unconventional path by taking our time to validate our convictions and building a suite of world-class products. None of this could have happened overnight—it has been a journey of patience and perseverance, and one that required staying true to our beliefs, through all the highs and lows. This campaign is a reflection of our approach towards sustainable growth. Watch video (English): youtu.be/4ivmul54L-w Watch video (Hindi): youtu.be/lJen7rbL6vg Website: zoho.com/time/ #GoZoho

would you take this job for $1,000,000/year?

Just as writing made memorising and reciting redundant, And printing made scribes redundant, So video and audio are slowly making writing redundant. As a writer that hurts. But does it matter in the end? Genuine question.

On the subject of AI and jobs: Hypothetically, if all software development were to be automated - I want to emphasize that we are _nowhere_ close to that goal - and all software engineers such as myself are out of work, it is not like human beings will have nothing to do. The problem is purely economic - how do people afford all the goods that pour out of automated factories that employ no workers? There are two possible solutions: 1. The price of all the robot made goods and all the AI-made and AI-supported software would drop massively and be close to zero or zero. Breathing air costs us zero and we don't complain about it! 2. The remaining things humans do, may get paid well - as an example, taking care of children, home cooked meals, nursing sick people, priests that minister to people, people who take care of soil health, water health, crop health and cattle health (we used to call them farmers), forest restoration specialists, local live performing musicians and so on may get paid much more and that circulates income widely enough for people to afford the goods pouring out of highly automated factories. To state it differently, IF (again, a big if) robots and AI automate all production work, the remaining work humans do like taking care of children would pay very well in terms of purchasing power of those now-ultra-cheap or free goods. This is fundamentally an economic distribution problem, a problem of political economy and not purely a technological problem. One key part is for governments to crack down on monopolies, particularly tech monopolies. Only that will ensure that the prices of goods reflect the very low cost of production arising from AI and automation. There will be at least one country in the world that would get the political economy right.

You can't actually make something safe without first making it. How do you make a car safe without making a car? The same is true of AI and anyone who tells you different has never built a thing in their life. We've been making things safe long before we ever understood them. Medieval engineers built soaring cathedrals that didn't fall down with only the basic math. As Bill Hammack wrote in The Things We Make: "To size the wall, the head mason would need to use a rule inherited from late antiquity, a rule that created the Pantheon and Hagia Sophia: a stable arch results when the wall’s thickness is a little more than a fifth of the arch’s span. The head mason, though, had likely never learned to read, let alone calculate a dimensional ratio. Instead, he carried out this rule without a yardstick marked with numerals, without Euclidean geometry—with only the most basic mathematics." We built things in the past long before we understood how they worked and it's always worked out perfectly. You can tell if something is working by the results. We made bread long before we understood what yeast was and we made cathedrals stand up and we fashioned o-rings without understanding what made them long lasting and resilient. We can and will do the same for AI.

A good cover letter is always the first gate for landing a job at 37signals. I don't care if we're the last company to require them, we're never going to hire someone who can't write a single, personal, fluent page on why they want THIS job. world.hey.com/jason/cover-le…

Amjad Masad says the real danger isn't killer robots. It's that AI gives us everything we want. An infinite pleasure machine. Simulated struggle and meaning. “You don't need to leave the world.” Fermi's Paradox: Maybe civilizations don’t die, they disappear into the simulation.