The Future of AI and Software Debate
Debate or Debacle? A Takedown of the Now Infamous Citrini Report.
Ok, here comes another rant post. So please bear with me. I have to get this all off my chest. The only way I know how to do that is to “Write It Down”. Your choice to read! Here goes:
There’s been a renewed and roiling debate happening in private and now public media of our world’s future with AI. It has recently surfaced and reflected in the public equity markets and how Wall Street is valuing software and SaaS stocks - what I consider the user interface and front door to our world’s technology stack. This AI fear has sent our software technology stocks down between 30-50% from their all time highs recently.
Specifically, there is a now report from Citrini Research that went viral in my personal network about 7 days ago. If you haven’t heard of it, you have now. I’ve posted the full report at the end of this post to document it for history and so you can read for yourself the insidious dystopian fear this report so eloquently created.
Over the last several months, I’ve been very steadfast with my own point of view of AI’s impact on the arc of our historical and future technology. More recently, I’ve been on my soapbox with anyone who will listen as to just how wrong Wall Street is with respect to declaring the software and SaaS industry effectively dead. With this post, I’m not shouting into the internet void.
Stay With Me… Don’t Leave Yet.
For those of you who know me very well, you understand I’m a huge Monty Python fan. People love my other Monty Python Zoom waiting room video of “You Wait Here” sketch.
If you are a geek like me and experimented with the 1,000s of Zoom advanced settings, down there buried after at least 100 clicks is the ability to use a video for people waiting to be let in on Zoom. Apparently few else has found this setting or actually uploaded a video for it. I did this 3+ years ago, and it’s the best ice breaker I’ve created for those waiting for me let them in.
The whole AI-software debacle reminds me of the famous “Bring Out Your Dead!” sketch… ”I’m not dead!” “Oh you will be in a minute!”
Bottom Line: Software and SaaS is not dead yet!
Along comes the Citrini Research Report last week, and I was once again fascinated by the psychological construct it used for maximum storytelling of fear and influence effect.
So much so, that I now feel compelled to expose these techniques and to write my own point by point counterpoint as I’m not seeing much of a counter -argument/debate (due to how effective Citrini’s psychological writing techniques were used). Think of any great Hollywood movie and how it strings you along to an “obvious conclusion” that anyone can see unfolding.
My goal is to forever document this moment of time and how my views differ significantly from the “world view” of how technology is unfolding before us as of early 2026. Note: It’s only been a mere 3+ years since OpenAi/ChatGTP burst onto the scene in November 2022.
As usual, I’ve also scoured the web to try to condense others “best ofs” when it comes to this subject to enhance my own thoughts and will source them accordingly.
Here is where AI is great. I’ve asked it to summarize this lengthy report into it’s key points so I don’t have to. In summary, Citrini’s memo narrates this horror script:
A storytelling memo set in the year 2028 in a post-mortem style depicting “what happened?”
How agentic AI causes a collapse of our entire economic system focused mainly on software and SaaS displacement of seat licenses, renewal revenue, and margins.
Then, broadening into an overall systemic economic income collapse in an accelerating loop without any natural brakes starting with job losses and cascading into consumer consumption, credit, and finally asset prices (led by home values).
Taken at face value the step-by step nature of this report allows anyone to “enter the story” and experience what is obviously going to transpire.
The problem, which is the problem with most stories and especially horror films, is the inability for the main cast of characters to change their fate. These stories usually always have a set of unrealistic assumptions that are hidden well and inherently false.
Here are 3 of Citrini’s key false assumptions:
Worldwide Economic demand is fixed
Productivity gains don’t expand markets nor do they create new companies and industries
Our Societal engines and human behaviors stay fixed and do not adapt
We know from our past that all 3 statements are overwhelmingly false. And yet, Citrini, uses these falsehoods to propel their fear based propaganda - “research”. A classic strawman fallacy.
Kobeissi’s letter has a great take on Citrini as well which I’ll summarize below:
Commoditization is Not Collapse
https://x.com/kobeissiletter/status/2026125293149163694?s=61&t=-YygOMqKQ790TaVVKGYKmQ
The PC revolution of the 80s commoditized computing
The Internet of the 90s commoditized information distribution worldwide
Mobile and “Cloud” commoditized information access anytime/anywhere
AI is now commoditizing intelligence
The doom loop (Citrini’s example) becomes dominant only if AI replaces labor without materially expanding demand. The optimistic scenario emerges if cheaper compute and productivity yields entirely new categories of consumption and economic activity.
SaaS is a software delivery mechanism, not the endpoint of value creation. Value creation is ease of use, speed, and convenience. All three will be enhanced by AI, not destroyed by it, and those companies with the data and their close customer relationships will win… not lose to newcomers who don’t have either the customers or the data or the specialized differentiation moats of what these customer truly want and how they behave.
The next generation of software is adaptive, agent-driven, outcome-based, and deeply integrated. The winners will not be static tool providers, they will be those who can best adapt to change.
Every technological shift reorders the stack, and the companies who only price static workflows WILL struggle. The companies who own data, trust, compute, energy, and verification will thrive.
Tariffs are tools of protection in a world where domestic industries struggle to compete on cost. But if AI collapses production costs everywhere, why would we need tariffs anymore? In a high-abundance environment, protectionism becomes economically inefficient.
AI amplifies outcomes. It can amplify fragility if institutions fail to adapt and it can also amplify prosperity if productivity outpaces disruption.
The Anthropic advances are signals that workflows are being repriced and cognitive labor is becoming cheaper - a clear transition.
But transition is not the same as collapse as every major technological revolution has looked destabilizing at the start.
The most underpriced possibility today is not dystopia, it’s abundance. AI may compress rents, reduce friction, and restructure labor markets, but it may also deliver the largest real productivity expansion in modern history.
The difference between “Global Intelligence Crisis” and “Global Intelligence Boom” is not capability, it is adaptation.
And the world has always found a way to adapt.
From Howard Marks newsletter (Oak Tree Capital):
Everything Claude or any other AI has learned has come from human-written text. It has no experiences, no embodied understanding of the world, no genuine comprehension.
I’ll add my own footnote here to Howard’s views highlighting my own views:
AI cannot see, hear, smell, taste, or touch. It cannot read body language and it does not experience or know how to deal with emotion. All of these are inherently unique human traits and AI will never be better at replicating them than we do. AI can only store a model of the world in its memory and pattern match against what is true today or the past. It has no better way of predicting the future than anyone else since the future is created by humans and fallible human decisions.
More Howard:
Everything it produces is ultimately some sophisticated rearrangement of patterns it absorbed from existing human work. It’s extraordinarily impressive pattern matching – maybe the most impressive pattern matching ever engineered – but it’s not thought. It’s not reasoning. It’s statistical recombination.
And if that’s true, then there’s a ceiling. It can remix what humans have already figured out, but it can’t break genuinely new ground. It’s a very talented cover band, not a composer.
Finally Marc Andreessen wrote his Techno-Optimist Manifesto back in 2023 and went so far as to post it as a takeover on a16z’s home page for a few weeks. I’m posting the link to this opposite Citrini’s report at the end of this post, for posterity sake. I’ve also asked my AI to do a side-by-side of the 2 posts just so everyone can see the now 2 most famous memos on the subject (2023 vs 2026; VC vs HedgeFund).
Now here’s my personal take as promised. I’ve grown up in the world where technology meets SMBs (small to medium sized businesses). My immediate reaction to the Citrini Research Report had many of Kobeissi’s and Howard’s arguments in my head before I found/sourced them. I’m very happy to have now written them down here and have the ability to add my own views.
Over the next 5-10 years, AI will become the great equalizer for SMBs, the vast majority of which are 2-25 employee firms. These companies have historically lagged their larger rivals due to the opposite of “too big to fail”… otherwise known as “too small to afford”.
AI advances are rapidly leveling the playing field and provides all these SMBs, including worldwide individuals, with access to information and workflows previously mostly inaccessible due to costs or local availability. They now almost immediately have been provided very cheap access to:
World-class research on demand
World-class software development with 1 person (no 10 person engineering org needed)
Automated workflows not requiring a “manager” or administrator
My belief is we are entering a new Era of “Community GDP”… the holy grail of the local economy level which powers nearly 1/2 (50%) of US GDP.
SMBs are 99% of all US Businesses
Over 35 million companies
46% of all US employment
And 44% of all US GDP
While these SMBs have historically lagged far behind the technology adoption curve, they are all about to catch up very, very quickly.
Few talk about or really understand the true power of democratized countries. The power lies at the mostly invisible local and individual home level. When these economic engines of productivity come to life and finally start competing on a more level playing field, the impact on the overall economy will be massive.
Your local HVAC company will get their jobs quoted faster, their inventory delivered faster and invoices and cash collected nearly automatically. This means less time spent administratively and more time spent on performing more jobs per week/month. More jobs = higher income for the company and the individual installer.
Wait, did I just say that AI will be a significant creator of job growth at the local, SMB level? Yes, I did. That’s exactly what has happened in every major technological shift from the printing press to the railroads to the automobile to computers. Jobs actually expand. New companies do things faster and create larger companies and eventually entirely new industries to service everything.
The Aha insight? New technology saves time. Time allows for more thinking time. More thinking time increases creativity. Creativity creates all new industries.
People and companies become significantly more productive and productivity solves our worldwide labor shortage problem. We’ve already forgotten that we’ve all been worrying and wringing our hands over the last decade of the impact of our labor shortages. Somehow we’ve been surprised of our continued low unemployment rates when the reality is that we don’t have enough people to fill most jobs.
While many are saying many jobs won’t be needed anymore, they don’t realize that we will retrain or the people will move to the jobs that are needed. That has always happened and will happen again.
Our economy is not fixed. We’ve never had fixed demand or fixed numbers or types of jobs no matter how much Citrini wants to to believe that.
Once the playing field is leveled and know-how, intelligence, and tasks become cheaper, our capacity constraints become ambition constraints.
Those who are the hungriest with the same tools will win.
Those who are the most creative or who have the best taste wins.
Those with specialized or custom knowledge or data wins
As good as AI is, it absolutely does not have taste, ambition, emotions, or custom knowledge. It only has historic data in it’s model… not new, creative data. I’m rather confident that nobody can train their LLM for taste, emotions, or unique cultural values either.
We must remind ourselves AI is simple pattern matching and reassembly of mostly Reddit and the internet at large. It’s a language model that takes words or text and reassembles them into much better structured words and text with higher precision and quality.
The winners of the exponential abundance of intelligence and workflows will be the same companies people TRUST to do business with who have the same level playing field advantages of the bigger companies.
The 12 person manufacturer in Tiffin, Ohio
The rural health clinic in Lyme, NH
The five person logistics company in Bardstown, KY
The 6 person marketing shop in Boise, ID
A new entrepreneur retailer with one new idea and one location who can now create a real brand and launch it into a new worldwide franchise opportunity.
AI has not centralized intelligence into the largest companies. It is open sourcing it and distributing it to the lowest levels of our economic playing field.
Citrini’s memo frames a doom loop of AI getting better - companies laying more and more people off - spending collapse and our economy breaks. Step by step this all seems plausible until you zoom out and realize the core assumptions here is everything is fixed. Demand is fixed. No new creation. No new wealth created by new ideas and new creativity.
The Citrini report doesn’t factor in the democratization of individuals and SMBs that powers nearly 50% of our economy which have historically been competing with one hand tied behind their backs.
For 100+ years, scale has been the scarce capability of large enterprises with large capital summed up by the phrase “It takes money to make money”. But what if it didn’t take money to make money? Hmmm… that’s interesting.
What if SMBs all of a sudden had a significantly increased:
Access to capital
Expensive systems became cheap
Talent wasn’t specialized and widely available
Distribution of products beyond local became faster and easier
That 10 person SMB can now:
Research at the same speed and quality of the 10,000 person corporation.
Code with the same speed and quality of a 100 person engineering team
Produce back office workflows that used to cost millions of dollars of software
Automate with AI Agents that used to require hiring lots of employees or consultants.
Citrini writes a line I totally disagree with. It’s simply not true. Not historically and not today.
“This is the first time in history the most productive asset in the economy has produced fewer, not more, jobs.”
Citrini’s statement again leads with the logical fallacy that everything is fixed and further does not admit where all jobs have ever come from.
The most productive asset (or technology improvements) in our entire human history have never “created jobs” directly.
They have only created the conditions for new ideas, new firms, new services, new categories, and new consumer demand.
Electricity didn’t hire people. It made factories possible. Refrigeration possible. Modern retail possible.
The Internet didn’t hire people. It made e-commerce, digital services, and global distribution possible.
AI is the same kind of general-purpose shift for Intelligence. We are simply freaking out because it is happening at the speed of a computer chip vs taking a decade or more to unfold.
To be clear: nobody’s framework fits. That does not mean we are doomed as Citrini suggests.
It simply means We Are Early.
Of course, we need and we will create new economic frameworks. The question is not whether we build them in time. The question is whether we point our new technology and the capabilities they afford at the right parts of our economy first:
I believe this power will be and should be pointed squarely at Main Street. Small town America. The 2–25 person engine. I grew up there. I know our rural economy better than most. Lots of my close friends still live there.
When technology and intelligence friction goes to zero (one of Citrini’s core paragraphs was titled “When Friction Goes To Zero”) in this very large segment of our economy, it won’t be a crisis… it will unleash a small business renaissance the likes of which we’ve only historically seen in a small pocket of the world called “Silicon Valley”.
Here’s my Bottom Line:
Every “inevitable AI forecast” is consistently getting wrong and forgetting that we humans adapt. We always have and we will adapt again this time. We aren’t a fixed set of puppets that will be eventually controlled by technology or robots. Markets will thus adapt… led by humans. Business models will adapt… led by humans. We will not cede control to the machine. It’s simply not our nature.
Here’s historical proof/facts and perspective of how these other new technologies in my lifetime were going to kill our economic growth.
The Internet did not “kill retail commerce”, it clearly expanded the TAM (Total Accessible Market) of spending and accelerated it.
FACT: 30 years after the birth of Amazon, E-commerce is still only around 20% of all retail commerce.
Mobile didn’t kill desktop. It expanded the overall usage of software.
Smartphones and their apps didn’t kill software jobs, they created a new app economy, mobile payment economy, and all new consumption behaviors.
Social networks didn’t just change marketing. They created entirely new distribution methods and 100s of billions of dollars of new business value.
25 years ago we didn’t have the likes of Google, Apple, NVIDIA, Amazon which are now each multi-trillion dollar companies.
That value didn’t come from Citrini’s and others who argue using the core assumption of “fixed demand” or “status quo consumers behaving the same way forever.”
Value has always been created from new products, new behaviors, new industries, and new jobs that we couldn’t predict until they were born.
The most insidious assumption inside the Citrini crisis narrative is the dystopian premise that human behavior stays fixed. Citrini fails to even acknowledge that new companies won’t be built with the new technologies. Or that new categories of intelligence will attack old moats. Or that current workers won’t move up and learn new roles and become the managers of the new AI tools. Citrini doesn’t acknowledge these truths since they would destroy their narrative.
No, we won’t be the same people with the same behaviors in 2028 as we are in 2026 with the new Agentic AI “scary software”. If that were true, we’d still be stuck in our 1930s Depression Era behaviors.
Finally, here’s the higher level point I needed to document that applies to not only Citrini’s memo but any other “future memo” that is always hard to debate.
There’s a reason this step-by-step future storytelling is so persuasive, and it’s because we can’t visualize the future, and we can only relate to today.
The psychology Citrini uses is:
Narrative Transportation: this technique reduces counter-arguing. It’s the classic Hollywood movie technique or campfire storytelling technique that “transports” you into the story by stitching together today’s facts and beliefs and allowing your to simulate the world with only today’s fixed mind set.
Simulation Heuristic: step-by-step using “and then… and then… and then” pulls you along the narrative you are transporting on and simulates inevitable fear. When that narrative does so very smoothly, it can feel very tight and the human brain typically fails to challenge some core assumptions on this narrative transport train that are false. Challenge those assumptions (such as fixed demand) and you can get off the simulation train. Daniel Kahneman and Amos Tversky, 2 very famous psychologists wrote the research on this psychological technique.
Fear: fear makes risk feel larger. When a great narrative or horror film produces a strong emotional signal, your logical brain stops thinking and fighting and starts flighting. This is the most insidious part of our animal brains for those that create narratives that prey upon our DNA this way.
The Future Is Impossible To Visualize: therefore, any alternatives are also hard to visualize. Therefore the human brain can only rely on the facts that are known today and how this new powerful tool impacts today’s facts.
These types of cinematic forecasts can feel “undebatable” since they deftly implant narratives and fears using today’s facts into our animal brains.
The Citrini Report tries to trap the future inside the status quo of today. That’s false. The future does not look like today. It never has and it never will.
When we transport ourselves to 2028, we will see new tools, improved lives, new companies, new economies, and individuals and SMBs so much better off than they were in 2026, it will be obvious… but we will only see these by looking back with 20/20 hindsight.
AI is simply the latest inflection step on the technology staircase we’ve been climbing for the last few centuries. We have delivered more capability and improved lives to more people in more places at every step.
Bottom line: Human progress is compounding and will not all of a sudden start degrading. Distributed intelligence is the next compounding value engine.
I remembered and added Mark’s Manifesto above and the quick summary below only about an hour ago right before this post was scheduled to go live. It turns out I’m very aligned with Mark’s view of the future. I also quickly asked my AI to write a comparison of both reports (it did a great job).
Citrini vs. pmarca (Mark Andreessen) — same headings, opposite worlds
1) Core thesis
Citrini: Abundant machine intelligence is a macro hazard—it unwinds the “intelligence premium,” breaks labor income, and pulls the economy into a destabilizing loop.
pmarca: Abundant intelligence is a macro flywheel—technology is the lever that turns scarcity into abundance and raises living standards over time.
2) The mechanism
Citrini: AI removes friction → destroys tollbooths/intermediaries → compresses LTV/margins → firms cut payroll → demand drops → firms buy more AI → loop accelerates.
pmarca: AI removes constraints → lowers cost of doing hard things → expands what’s possible → new products/categories/firms form → demand expands → capability diffuses → compounding progress.
3) Human behavior assumption (the hidden hinge)
Citrini: People and institutions behave mostly like today, but with better software; adaptation is slow; new industry formation doesn’t offset displacement fast enough.
pmarca: Humans adapt fast when tools improve; entrepreneurs create new industries; markets reorganize; society composes new “jobs-to-be-done” around new capabilities.
4) Jobs + wages
Citrini: The productive asset (intelligence) becomes abundant and produces fewer jobs—demand falls because payroll shrinks.
pmarca: “Lump of labor” thinking is wrong; productivity shifts work up the stack and expands total value creation (new roles, new firms, new demand).
5) Markets + moats
Citrini: Friction moats collapse; habit economics die; platform margins compress.
pmarca: Bad moats die; good moats (trust, reliability, brand, relationships, execution) strengthen.
6) Distribution rails (where the value actually lands)
Citrini: Intelligence centralizes into agents that optimize consumption—pushing prices down and destabilizing incumbents, with broad collateral damage.
pmarca: Every general-purpose tech becomes a rail; rails push capability to the edge (local + personal), unlocking “wealth of technology” for billions to participate in the global economy.
7) The punchline
Citrini: The risk is a self-reinforcing contraction—productivity without paychecks.
pmarca: The opportunity is compounding improvement—abundance that makes life better at scale.
This post was long enough and I cut out a bunch that I saved for Part II next week. Thanks for reading this far and I’ll hopefully see you here next week!
