#10: AI Armageddon: Ruining Lives Or Curing Cancer? Depends on Your Perspective

TLDR

Recent AI happenings (see below) have sent shockwaves through capital and labor markets as well as enterprises over the past 2-3 weeks. Fact: AI is changing what is possible at a pace of change we’ve never experienced before. Investors are panicked. Start-ups are rethinking their product and business strategies. Employees are clinging to their jobs. A few thoughts to create context for the panic and counterbalance our fears: 

1. We live in a physical world augmented by digital services - not the other way around. The fact that anyone can build any digital service or experience anytime with no coding skills is becoming real. However, we still live in a world where healthcare, judicial systems, drug trials, regulatory and legislative environments are analog. Just because a tool can replicate the work of an attorney doesn’t mean the court system can process it more quickly or speed up trials.
   

2. Consumer behavior takes time to catch up to the possibilities of technology. The breadth and sophistication of digital experiences far exceeds consumers’ ability to consume it. Just because one can use an online conversational interface to access customer care, doesn’t mean they want to do so. And even if a consumer is willing to converse with a chatbot to get service, it doesn’t translate into their interest or comfort with making a purchase through a chatbot. Consumers also glean satisfaction and confidence from learning to do new things and completing tasks. Not everything has to be fast. This list is also long.
   

3. Consumers lean into experiences they enjoy. For example, consumers enjoy shopping. More than 80% of U.S. retail spend is still in store. Given the prevalence of agentic commerce trials in grocery shopping, there is clearly a subset of consumers who don’t enjoy shopping for food. While it may be possible to place an order and pay for groceries within the OpenAI environment (and have a $7,000 Korean refrigerator monitor the inventory in one’s fridge), most consumers will still go to the store because they want that experience. They also still enjoy eating food. The market for protein bars has not overtaken the food or restaurant industry. The list is long.
   

4. Critical thinking skills will be increasingly important. These AI tools that do so much good will also enable nefarious events - the least of which is that anyone can create and circulate (obvious) deepfake content or services. Consumers in the US lose between $20B and $200B annually to fraud or phishing schemes. (Think of the movies, Beekeeper or Thelma.) Many of these are perpetrated by getting an elderly citizen on the phone, chatting, and building trust. Now imagine that a voice bot could do so at scale.
   

5. The makeup of employees at companies will change, and we aren’t yet sure how. AI offers unprecedented compute power with autonomy at a speed we’ve never seen before. One might use agriculture as an analogy.  In 1850, 4.9M Americans or 63.6% of the population worked in farming. The numbers peaked in the early 1900s before declining to just 1.5% of our workforce today. That happened over 175 years so we had time to adjust while building new skills and industries. AI is progressing faster than most of us - personally or professionally -  can process the impact. We may see a shift of similar impact in less than a decade. 

What it means (WIM) for digital experience leaders? 

1. Challenge your assumptions about the talent, time, and financial resources to execute your goals to roll out new features or versions of your digital experiences.
   

2. Keep using the tools. If the models failed to deliver results (e.g., analysis, PPT slides, spreadsheets) 2 weeks or 2 months ago, try again. The tools are improving quickly. Think of the most talented young associate you ever hired and how fast they learned.
   

3. Reassess your “buy vs. build” decisions. You may find yourself doing more internal development than you were one to two years ago.
   

4. Reconsider how you pay for consulting and the value you expect. Also consider, whether or not you want to require disclosure of use of AI tools - what, when, how, and where along with what has been verified. You can ask questions of ChatGPT, Gemini, and Perplexity. All of these tools can produce research reports. Know what you truly value whether it is expertise, judgment, or personal networks. 

AI innovations are having a rapid and dramatic impact on markets, businesses, and consumers’ mental health. Here’s a quick overview of the most exciting or unsettling news since late January /  early February: 

• OpenClaw (known until recently as Moltbook) hit the news when AI agents started posting and interacting with one another while humans observed them. The agents created a new religion (Crustafarianism), complained about humans, and encrypted messages to avoid human observation. OpenClaw is marketed as a personal AI assistant (e.g., set smart reminders, manage subscriptions, find recipes, automate data entry). It is almost like a social network for bots. The Guardian has a good article explaining what it is and does. What it means (WIM): If you hand over control of your personal data (e.g., health, financial, email), account credentials, payment credentials, and permission to act, it’s hard to say what might happen.
  

• Large public tech companies (e.g., Alphabet, Amazon, Meta, and Microsoft) posted earnings and planned capital spend that when added together represents about 2.1% of the U.S. GDP in 2026. The WSJ had a great graphic and headline, “The Cost of AI Efforts Tops Moon Landing.” According to the WSJ, only the Louisiana Purchase cost more in comparison. Order of magnitude, this is about six times what they spent in 2020. Think back to 2021 - the year that Mark Zuckerberg went all in on the metaverse and renamed his company. Feels like an eternity. And, this $670B is about twice what they spent in 2025. And, no, their revenue is not growing at the same rate. They are building and stockpiling compute power.  Investors are proceeding with caution because there are not common mechanisms for depreciation of these assets. WIM: A handful of companies plus the privately funded ones (e.g., Anthropic, OpenAI) will own a high percentage of the future infrastructure.
  

• First (and mostly) Anthropic and then OpenAI released new agentic coding models to developers. Anthropic released Claude Opus 4.6 on February 5th while OpenAI released GPT-5.3-Codex on February 12th. Stepping aside from the tech details and simplifying for a business audience, these upgraded tools shift agentic AI’s role from collaborator to autonomous teammates capable of handling far more complex and time-consuming tasks. They code, plan, debug, and deploy code more autonomously at faster speeds. While I do live in a bit of a bubble in SF, hardly a day goes by without someone boasting to me about the new app or web-based tool they built in just two hours. WIM: With a little persistence, just about anyone can build anything. That doesn’t mean it will have an award-winning UI, protect anyone’s privacy, or be the most efficient tool.
  

• Investors began selling off enterprise SaaS stocks when Anthropic’s new coding model launched on February 2nd. The sell-off has continued through OpenAI’s release of its coding model. Depending on the day of the month or the analysis, between $300B and $1T in enterprise SaaS market cap is gone. Last year, the popularity and promise of vibe coding gave us all an inkling of what agentic coding tools could do: build anything digital we expressed in natural language. When Anthropic launched Claude Opus 4.6, investors seemed to decide that anyone could build supply chain, CRM, marketing, contact center, human resources, financial reporting solutions and more i.e., the market for enterprise SaaS would collapse. WIM:  There is a difference between possible and plausible. While I don’t want to underestimate or undersell the potential of coding tools that allow folks to build digital experiences or automation far more quickly and easily, there is more to the equation. We could all grow our own food, but we don’t. Companies need to focus on their core competencies whether it is building engines or developing and refining cures for cancer. On one hand, they have installed solutions that are core to their daily operations. On the other, they may not want to “invent the wheel” on functionality, services, or features for overhead functions outside of their core expertise.