HIVEMIND I: AI Market Commentary & Predictions

In this new series called HIVEMIND, we will delve into various technology, market, and macro predictions and commentary. Here we focus on recent AI news, dispelling broader AI narratives, and sharing overall predictions in the market.

Off the back of the sensational headlines and reactions to Jack Dorsey’s company Block laying off 40% of its team “due to AI”, I wanted to share some burning thoughts and predictions on the AI market. The first thing I jumped to when I saw this headline was to look at Block's share price. Down 80% over 5 years after peaking in the market mania during COVID. Then I looked at their employee count prior to layoffs: 10,000+. Then I looked at 2025 earnings, and saw 3 bad misses and 1 modest beat. Then I read about other layoffs they had started rolling out in 2024. It became clear this was a planned, overdue, and necessary multi-year restructuring in a situation where bloat, over-hiring, tough fintech margins, and 'shareholder value' were the main drivers, not AI in the slightest.

The AI job loss dynamic in the market is misunderstood and will be used as a scapegoat for years to come. We've been at 4.3% unemployment for 12+ months, with more automation than ever, but more importantly with many organizations still holding onto excessive bloat from a previous inflationary phase in the market. Those are the cases where job loss like this may occur. Block is an awful example of "AI-caused jobs loss" and as more companies carry out shareholder and stock-benefitting layoffs with AI as the supposed catalyst, look no further than the metrics mentioned above to see if it's real or if the company had other structural reasons to inevitably cut jobs.

We don't deny that specific sectors and roles in the market will be impacted by AI, namely entry-level analyst, design, or research-oriented positions to name a few. For the existing workforce and other roles there ia more nuance around retraining, reallocation, or remaining untouched or benefitting based on job function, but many of those going into certain entry roles will need to pivot. There will be many second order effects from this however, and this will start to become alleviated by younger people specializing in other areas with help from parental, peer, online, and academic sentiments. Ironically, we've had the post-college job difficulty for a long time due to colleges marketing useless majors to students and burying them in debt. Now we suddenly seem to care about post-college job difficulty when it impacts things like junior finance jobs. The reality is, people and the market will adapt to this and plenty of other areas will see increased demand in parallel and we argue that nets out and goes positive.

AI doomsday predictions rely on everything happening in a one dimensional vacuum with no agency or free will, and misunderstand or invert incentives in the market. Doomsayers completely ignore second order effects, human action, adaptation, and many basic economic principles, along with the history of automation up until now. In the case of technological history, everything is relative. Even if analogies like the steam engine, tractors, electrification, calculators, ATMs, spreadsheets, and the internet itself are seemingly cliché and all have their own nuances, they hold up the core principle that fears around innovative breakthroughs have never played out the way people predict. In all of these examples, some specific/specialized displacement occurred over time, but nowhere near what was predicted. The market adapted and all measurable economic metrics, including labor demand itself in these industries, ended up increasing.

AI "feels different" than these examples because we've come so far technologically and we are living this cycle in the present with many fearful narratives being presented, all amplified by social media. The past is simple to write-off as different in hindsight and we often forget the concept of relativity when it comes to history. In its current state, the public views AI as more futuristic than it really is and the Sci-Fi brain or Black Mirror watcher will automatically assume the worst forms of dystopian accelerationism are around the corner. They also assume the worst intentions, that companies are evil, or that runaway AI is about to not only take jobs and ruin the economy, but end the world in some Hollywood way. We would take the opposite side on those kinds of predictions all day.

When it come to technology, pessimists and doomsayers have been wrong throughout history and they are wrong again. We've been raised to largely be pessimistic and cynical about technology and completely take the value it has created for granted (this goes for energy infrastructure and fossil fuels too). Mass media has always misunderstood and distorted innovation. Labor markets have always lagged technological innovation. And the doomsayer archetype and luddite have always existed throughout history to chase a sense of purpose. It's nothing new and it's not different this time.

Beyond the technology itself and speaking to the market and valuations, the hype cycles this fund has experienced in crypto since 2015 have given perspective on market psychology and technology adoption. It's certainly not apples to apples, as AI is creating much clearer value and has more applicability than crypto ever did across the economy, but human psychology is the same in the market. Despite being bullish investors in AI companies for over 10 years, today there is an exaggeration on how AI is impacting enterprises and the economy both positively and negatively and a misunderstanding of how it will impact the economy over time. Valuations are beginning to show signs of initial fatigue in som areas but likely have more room to run than people may imagine. With that said, we're well beyond the first or second inning.

Below is a list of predictions, some touched on above. The main thrust is we are highly optimistic on the long term potential of AI, with some mid-term excitement but many nearer term doubts during a seemingly growing manic phase (from both valuation and media coverage perspectives). We are most excited for some of this noise to quiet down so real builders, business models, and winners can be defined. There are many more optimistic predictions that the fund has around specific use-cases or application and excitement around our current portfolio companies, but most of the predictions I wanted to write about today are related to the market, macro-related topics, or various doomsday scenarios.

-Over the next 5-10 years, AI will not cause net job loss. Specific pockets of organizations or sectors may be impacted, and specific entry levels will have a hard time landing certain roles, but this will be made up for in other areas in the economy. You will see plenty of scare headlines, but macro jobs numbers won't see substantial instability.

-AI will continue being used as an excuse for layoffs that needed to be done due to corporate bloat and over-hiring during the COVID/zero interest rate/inflationary era. In that sense, "AI-caused Job Loss" will be a mirage.

-Unemployment (currently at 4.3% as of writing this) will remain structurally low around 4-6% over the next 5-10 years (barring any outliers/black swan events in markets, geopolitics, etc). If there is an increase beyond that, it will be provably macro-related, not due to AI.

-Software-related spend and hiring has grown for 50 years despite massive productivity and efficiency gains, and will continue to grow even with AI. More products, new markets, new teams/divisions, more code shipped -> more QA -> more maintenance, etc. Most companies will not prioritize shrinking teams, they will increase ambition and consumer/enterprise expectations will evolve.

-The current AI rally will go on longer than people think, but will have a corrective phase in the coming years. It may not be dot-com level, but valuations will be challenged and temporarily contract. The early warning signs are there. Smaller AI companies will fail, untenable business models will be exposed, and executives will re-think their AI-related R&D spend. The AI hysteria will quiet down and there will be a 'slowing' period in both valuations and adoption. Innovation will still be happening, just more quietly until additional breakthroughs and adoption occur and the market resets. This is the period where real winners will be defined.

-Despite potential ups and downs between, NASDAQ hits ~35,000 by 2030 and ~55,000 by 2040.

-As AI improves efficiency, demand for AI-related services and labor net increase in parallel (Jevon's paradox in action).

-The market is largely regulating itself when it comes to AI with internal checks and balances, security, and third-party solutions. Rogue or runaway AI is a doomer fantasy. Every company building AI is 100% incentivized against this happening for monetary and reputational reasons. There are other reactive forces in the market around cybersecurity and protecting against this outcome that are hardened and highly credible.

-AI adoption is iterative and adaptive over time, not disruptive in most cases. While new companies with AI DNA may begin hiring more selectively at the start or be able to scale with very lean teams, mid-market and enterprise companies with large client bases and network effects will act differently to compete with one another.

-Trust layers and network effects will be massively under-indexed when it comes to AI displacing products. A perfect example of this is claims like Perplexity's new research tool disrupting Bloomberg Terminal, when the value users are getting from Bloomberg are unique, nuanced, and network effect based. There will be a reality check around this, as hundreds of vibe-coded AI "replacement" attempts flop (GTM is also hard).

-AI Safetyism in its current ideological state largely goes by the wayside as more of the public understands and interface with the technology and the hysteria fades. More true in the US and Asia over Europe. Some politicians will still cling to the narrative. Companies who launched under the current safetyism-first banner and who don't pivot will flop.