Author: TT3LABS, Web3/AI/SaaS remote hiring platform
On February 26, 2026, fintech giant Block announced the layoff of over 4,000 employees, reducing its team size from more than 10,000 to fewer than 6,000. In a letter to shareholders, CEO Jack Dorsey stated:
Smart tools have changed what it means to create and run a company... a significantly smaller team, using the tools we're building, can do more and do it better.
Dorsey also gave his extremely blunt prediction:
I believe most companies are already behind. Within the next year, most companies will reach the same conclusion and make similar structural adjustments.
After hours trading that day, Block's stock price surged over 20%. This is the capital market’s tangible response: paying for the company’s AI leverage and efficiency.
An ordinary person with no programming background can now independently launch a fully functional app in a single night, thanks to large models. The capital market will inevitably raise a sharp question: What is the remaining value of the massive labor costs incurred by tech giants that employ tens of thousands of programmers to maintain the daily operations of a super app?
The trend of replacing human labor with AI will inevitably attract more large companies to follow. Anxiety is natural, but worrying alone won’t help. We must start by understanding the broader shifts in our environment and gradually work down to individual survival strategies.
AI is not just a tool—it is becoming a means of production.
Some people in the market are now using "Web4" to define the current stage. To clarify the timeline, let’s first outline the different phases of the internet’s evolution:
Web2
The core is the interaction between software and people, with different platforms using algorithms to capture user attention, essentially a battle for traffic.
Web3
Attempting to address the issues of digital asset ownership and value distribution. Many people simplistically equate it with cryptocurrency, but fundamentally, it still remains within the realm of博弈 over wealth distribution rules and has not touched upon the "production and manufacturing" relationships of digital products.
The Eve of Web4
AI has for the first time touched the very essence of changing production relations. It is no longer merely a tool for improving efficiency, but is becoming a new form of means of production. Those who use it more effectively can elevate their output by an order of magnitude.
Traditional team collaboration involves significant hidden costs: the judgment and industry intuition of excellent leaders are difficult to replicate for subordinates, and misunderstandings and rework losses are inevitable in multi-person execution. These are the "hidden taxes" of organizational operations, for which there were previously no clear solutions. AI has dramatically reduced this hidden tax—it has no learning curve, can execute high-quality tasks with clear prompts, and can simultaneously process multiple task streams. By combining an individual’s strategic judgment with AI’s execution leverage, one person can now achieve the output of an entire past team.
Of course, AI still occasionally "speaks with complete confidence while being utterly wrong," which means human review and judgment remain essential. However, model reliability is improving on a monthly basis, and the buffer window for purely execution-based roles is much shorter than most people realize.
Efficiency Equity and Deep Crisis: When Entry Barriers Are Eliminated
In the short term, ordinary individuals can gain an efficiency advantage by adopting AI tools. However, looking further ahead, once AI eliminates basic efficiency gaps and significantly lowers barriers to entry for professional roles, companies will realize that if overall business scale does not expand proportionally after individual productivity surges, maintaining the original workforce size becomes a liability.
Look at the current salary disparities. According to TT3LABS’s job monitoring data, since 2025, the AI job market has repeatedly seen compensation packages exceeding tens of millions of U.S. dollars, and these candidates are young AI engineers without extensive "team management skills." When Meta recruited core researchers from OpenAI, the signing bonus alone exceeded $100 million; the average equity compensation for OpenAI employees reached $1.5 million, and Anthropic’s senior research engineers can earn a base salary of up to $690,000 annually (excluding equity).
The money spent on capital is an investment in a scarce capability: making AI itself stronger. Those who can drive the evolution of foundational models see their value amplified geometrically across the entire business network. Others, whose work can be replaced by AI at lower cost, may see their valuation shrink.
This also triggers a deeper, underlying crisis: more and more people now instinctively turn to AI for answers, bypassing the essential intermediate steps of reasoning, verifying, and experimenting. Over time, this erodes their ability to think critically. The problem is that these "grueling efforts" are precisely what cultivate your intuition for problems. If you consistently rely on AI to replace this process, your role in the workplace will degrade into that of a "requirements translator": converting others’ requests into AI inputs, then passing along the AI’s outputs. But this intermediary step is exactly what the next generation of AI can most easily bypass.
Impact Map: Where do you stand?
Fear without coordinates is merely anxiety. Before discussing strategies, we need to first create a "shock map." This isn't about spreading panic, but about helping everyone locate their position.
Positions where high-risk tasks can be clearly defined by instructions
Entry-level coding, basic data analysis, standardized report generation, template-based design, routine translation and proofreading. These roles share a common trait: their work can be clearly broken down into "input → processing → output." A significant portion of the more than 4,000 employees laid off by Block fell into this category. Their professional skills were not lacking, but the tasks they performed were precisely those that large models can handle.
A standard worth asking yourself: If all your job responsibilities can be written as a single AI instruction, then machines are already capable of replacing you—it’s only a matter of when the company makes that decision.
Experienced mid-level traders are seeing the consolidation being "compressed."
Project managers, operations leads, and mid-level engineers. Their work involves judgment and coordination, which AI cannot easily replace in the short term—but it is being "compressed." Previously, five middle managers were needed to oversee different segments of a business chain and align with each other; now, AI has taken over execution upstream and downstream, allowing one or two people to run the entire workflow.
This group is facing a situation where "there are fewer positions available." Your skills haven't declined, but market demand for your role has dropped sharply. The way forward for this group is to leverage AI to amplify execution at the lower level, and to gain the authority to define problems at the higher level.
Master of Uncertainty in Value Appreciation
There is a category of work where the core is not about "doing things right," but about making decisions amid incomplete information and taking responsibility for the consequences—complex business negotiations, crisis public relations, cross-cultural organizational management, and high-risk investment judgments. AI can provide analysis and recommendations, but it cannot sign on your behalf, take the blame for you, or read between the lines of a glance across the dinner table to understand underlying interests.
These roles not only retain their value but also gain increased leverage, as the underlying execution costs have been significantly reduced by AI, allowing the same budget to drive larger projects and giving decision-makers greater reach.
In reality, many people’s jobs span more than one tier. A simple self-assessment: think about your daily tasks—how many can be clearly defined by a set of instructions, and how many require you to make decisions in ambiguity? The higher the proportion of the former, the more urgently you need to make a change.
Stop tool anxiety and turn public computing power into private advantages.
At the end of January, OpenClaw ("Little Crab") emerged suddenly, surpassing 170,000 GitHub stars within days. Model providers quickly followed suit: Alibaba Cloud launched one-click deployment, Tencent released CoPaw as a counterpart, and MiniMax and Kimi also introduced their own compatible solutions.
Then you’ll notice an interesting phenomenon: many people spend more time this month researching “how to deploy crayfish” and comparing “which meal plan is more cost-effective” than they do actually using AI to generate business outcomes. Everyone is chasing tools, but once you’ve set up your configuration, someone else can replicate it exactly in just two hours.
All large language models—OpenAI, Anthropic, Meta, Google, xAI—are trained on the same publicly available internet data. So they are essentially the same, which is why they are being commoditized at an extremely rapid pace.
— Larry Ellison, Oracle Q2 FY2026 Earnings Call
Conversely, if your work relies solely on the public capabilities of general-purpose large models, your output will be homogeneous—no matter how sophisticated your prompts are, there is no moat.
The real barrier lies in moving from public to private.
There is now a clear trend: more organizations, from large enterprises to startup teams, are deploying localized private models. The immediate reason is information security—no one wants to hand over core business data to third-party APIs. But this trend has an underappreciated ripple effect: as key players in the industry lock their data and knowledge behind private deployments, the amount of industry-specific information available on the public web for general-purpose models to learn from is decreasing and becoming increasingly outdated. While AI appears to lower knowledge barriers for everyone, the most valuable layer of industry expertise is rapidly disappearing from the public web and sinking into individual private knowledge bases.
So, the industry's "tacit knowledge" you've accumulated over the years isn't depreciating—it's appreciating, provided you put it into use.
Organize the non-standardized business insights scattered across your thoughts, chat logs, and historical emails into "context" that your private model can digest. TT3LABS backend data shows that candidates with over two years of Web3 experience have a significantly higher pass rate in initial screening than technical talent from big tech companies without industry background—the core reason being that industry-specific know-how carries far greater weight than general technical skills. Someone with three years of CEX operations experience understands compliance logic and listing unwritten rules; someone who has gone through two DAO governance cycles can judge proposal design and community sentiment inflection points; someone deeply immersed in niche content has an intuitive grasp of audience psychology and narrative pacing—none of these exist in any public training data.
Once you structure these private insights and integrate them into the model, your AI will no longer be a general encyclopedia—it will become a dedicated partner that works solely for you and understands only your niche. This depth of output is something others cannot match, even if they use the same general-purpose model.
The core logic is simple: AI outperforms everyone in processing public knowledge, but relies entirely on your input for private experience. Those who can combine deep industry know-how with AI are the key assets in this new division of labor.
Your knowledge base is the real "model".
AI models are evolving rapidly; today’s GPT, Claude, and Gemini may all be replaced by stronger versions in six months. But for you, switching to a stronger model is just a matter of changing an API endpoint. What truly won’t be replaced or outdated is the private data and knowledge base you feed into it.
Models are general infrastructure that anyone can use. But the industry insights, business judgments, and hard-earned lessons you feed into them are your exclusive "training corpus." The stronger the AI, the better it becomes at digesting your corpus—and the higher your private moat grows. So don’t worry about whether building a knowledge base now will quickly become outdated; your knowledge base is the one asset that never depreciates with model upgrades. As models evolve, your data moat only increases in value alongside AI’s growing capabilities.
Meanwhile, the traditional logic of workplace competition is being rewritten. Previously, employees could demonstrate dedication by working late hours, but machines operate 24/7, rendering all strategies based on "I can work longer than others" obsolete in the face of AI.
Many people say: "I still provide emotional value within the team." True, this is a uniquely human ability, but its premium depends on your level. When a frontline team shrinks from ten people to two people plus a row of AI agents, the role of "team lubricant" loses its context. Yet at the decision-making level, deep human connections—essential for complex business negotiations, high-stakes trust-building, and mediating conflicts across competing interests—become even more valuable as underlying costs decrease. Emotional value isn’t disappearing; it’s migrating upward.
Ultimately, what individuals should invest in most in the AI era is not learning which tools to use, but continuously cultivating the private AI that only you possess. Tools evolve; your knowledge base does not.
Three actions you can start right now
Returning to Block’s case, some were laid off while others stayed— the difference lies in who remains indispensable once AI becomes a standard production tool. Don’t wait for your company to provide AI training; starting today, we can begin taking these actions:
01.Shift from doing everything yourself to building workflows
The biggest trap for workers is using AI to "take shortcuts" (like using AI to write weekly reports or polish emails)—this is still a mindset focused on execution. What you truly need to do is see yourself as a "contractor," and restructure the core output of your current role into an automated AI production line.
Don’t try out a dozen new models at once. Pick one currently the most mature tool—like ChatGPT Plus or Claude—and force it to step into the most time-consuming, experience-heavy part of your workflow. Transform your original linear process of “manually collect data → analyze and compare → output conclusions” into “set up automated data collection → feed into an AI analysis framework → manually intervene to refine.” When you can use this workflow to compress a task that once took a week into just one day—with consistently high quality—you’re no longer just a single computing node; you’ve become a high-leverage “micro-company” yourself.
02.Transform your tacit experience into your personal digital avatar
Large models learn from public data—they understand all the theories—but they absolutely don’t know the hidden preferences of your company’s extremely difficult major client, or the landmines to avoid when coordinating with the finance department. These “dark knowledge” insights, gained through countless mistakes you’ve made, are your most valuable asset.
But these assets won’t generate compound returns if they only stay in your mind. Your current task is to use the customization features now available in large models—such as Custom GPTs or Claude Projects—to turn your experience into its “system preset instructions.” Feed it all your edge cases, failed post-mortems, and unspoken industry norms. Your goal isn’t to build a static knowledge base or notebook, but to “tame” a digital assistant that embodies your distinct personal business style and works exclusively for you, 24/7. When your digital twin is fully formed, no one using generic AI will be able to compete with you.
03.Strengthen your ability to define problems and your sense of responsibility
In your team, start deliberately practicing by handing over the work of "finding answers" to machines, while keeping the power to "ask questions" and "make decisions" in your own hands. AI is a perfect answer engine, but it can never detect the true business motivations behind a need. When your boss says, "I want to create a new retention strategy," AI will instantly deliver ten growth hacking theoretical models. But only you can combine current budget and development resources to point out, "Option B is perfect but not feasible right now; Option C, with half the features cut, best matches our current pace."
At the same time, you must understand one thing: AI cannot go to jail or take responsibility. When a company pays you a high salary, it is often buying your commitment to "guarantee" the business outcome. When you submit code or solutions generated by AI, you must have the confidence to say: "I have reviewed the AI’s output using my professional expertise, and I take full responsibility for the final implementation." This willingness to make decisions in ambiguous situations and bear the ultimate business consequences—the "premium for accountability"—is something machines can never replace, in any era.
Dorsey said, "Most companies are already late." But for individuals, the opposite is also true: most people haven’t even started preparing or recognized this trend.
Not everyone needs to become an AI expert. But everyone needs to ask themselves this question: In your work, which parts will machines eventually be able to do, and which parts are uniquely yours? Then, shift your time and energy from the former to the latter.
If one day AI surpasses humans across all fields—perhaps by 2027, perhaps by 2030—this is not a change you can watch from the sidelines.
It doesn't wait for you to be ready.