AI is after us all… or is it? The human brain is the only thing that can solve differential equations while running on pizza and chicken McNuggets, but LLM’s can shave off costs on tedious and mundane tasks.
As of 2026 the elephant in the AI room is rising costs:
Ex-Morgan Stanley and Galaxy Digital managing director Andrew Isaacs: “Token use will only grow, Goldman Sachs forecaststhat by 2030 we can expect 120 quadrillion tokens used per month. We’re only seeing the dawn of an non-custodial agentic economy: costly and far from perfect at the start, but cheaper and smarter as we optimize it.”
Not all tasks are equally automatable.
- As per MIT research from 2024 only 23% of tasks benefit from automation, while 77% of the tasks are better left to human operators.
- Fortune’s reprint of the research arrives at the same conclusion: “when a job requires multiple steps, creativity, or precision, AI replacements are more likely to fail than succeed.”
AI excels in data-heavy tasks:
- Healthcare, manufacturing, customer service and finance — especially in mundane parts;
- Other uses include tackling some of the tech debt code, faster time-to-market prototyping, quicker ways to crunch through vast sets of data, and an agentic economy where you can use AI agents to gain edge in financial analysis and operations.
What do smartest tech and finance CEOs have to say about AI?
LLM’s, or AI as marketers love to call them, are becoming a basic tool for everyday use: it’s not that you will be left behind, but your life will be easier if you learn to use it.
Some of the smartest people in the tech world already know why they need AI and for what.
Gabe Newel, CEO of Steam “monopolist” of half-trillion gaming industry
“Figure out how to use AI to do anything better. If you wanna be an accountant, learn AI, if you want to be an attorney, learn AI, if you’re gonna be a programmer, you know, make sure that you know how to use the tools that are comfortable with those programming concepts.“
AI Bias:
- Steam explicitly labels AI use in games and remained adamant to continue this practice even after facing criticism.
Larry Fink, CEO of BlackRock
CEO of BlackRock considers “there’s a real risk artificial intelligence could widen wealth inequality if ownership does not broaden alongside it,” in the 2026 Annual Chairman’s Letter to investors, “The companies with the data, infrastructure, and capital to deploy AI at scale are positioned to benefit disproportionately.“
AI Bias:
- BlackRock’s AI workflows are not a fling either — they secured $2.5 trillion of assets under management more for a total AUM of $10.5 trillion without increasing the amount of staff via their Aladdin portfolio management platform, which packs ML, NLP and proprietary models to manage funds.
Changpeng Zhao (CEO Binance / YZi Labs)
CZ seems to arrive at a similar conclusion: AI agents will be transacting in crypto because it’s easier: “The currency for AI is crypto, AI is not going to swipe a card, get an SMS code, that doesn’t work for AI. It’s going to be crypto.”
AI Bias:
- Changpeng Zhao points out that shovels of the AI race are the top picks for investment: “Compared to directly investing in AI itself, I prefer to invest in the ‘shovels’ that support AI, including data centers, power, and other computing infrastructure,” and BlackStone’s report confirms it;
- Binance’s ~$10B AUM YZi Labs doubles down on CZ’s words and invests in Vana, AI data sovereignty L1 blockchain that gives users control over their data;
- Pools resources to launch Giggle Academy for kids in struggling countries to get education from 1st to 12th grade, and if successful, get a job from employers who can search students based on their in-accademy grades.
Ex-Morgan Stanley and Galaxy Digital managing director Andrew Isaacs, Neyro COO
Andrew doubles down on CZ-style vision about agentic trading and BlackRock’s AI approach: “The future of trading is non-custodial, multi-platform, multimarketed and agentic. Crypto has workflows which are too cumbersome for manual analysis with opportunities on the edge of what’s relevant, an agentic approach can reduce mundane tasks and decrease informational load overhead.“
AI Bias:
- Neyro helps traders who use Hyperliquid to run custom trading agents without direct control of user assets or centralization of users’ strategies, while keeping user funds intact and abstracted from the agent;
- Agentic trading is in full swing, with Ethereum, Polygon, Base and BNB Smart Chain accounting for almost 111k agents in service;
- Non-custodial trading agents are responsible for ~63 million transactions and ~$7.5M USDC total volume in December 2025.
Arthur Hayes, Maelstrom and co-founder of BitMEX
Like a real bear who bets on market falls, Arthur Hayes (Maelstrom, BitMEX co-founder) in his essay “This is Fine” warned that if AI displaced 20% of all white-collar workers it would result in 13% equity drawdown, or around $1.1 trillion in losses, which can result in recurrence of 2023 regional bank wipeout, and FED-style money printing to cope with losses similar to 2008th crisis.
And Hayes is heavily biased into the AI narrative:
- backed GenLayer, a natural language processing layer platform, via Maelstrom;
Bottom line — people remain relevant, yet it would be prudent to prepare for a more agent-driven economy in the future as a contingency.
Source link
#coming #mundane #workflows #Tech #CEOs #proper
