What Gets Measured, AI Will Automate
AI doesn't need a sci-fi upgrade to upend the economy—current models, and the cheaper, more capable versions already in the pipeline, are set to disrupt nearly every corner of the labor market. Their surprising performance across text, image, and video threatens to upend how work is done across the creative ranks of writers, designers, photographers, architects, animators, and brand advertisers, as well as the spreadsheet crowd of financial analysts, consultants, accountants, and tax preparers. Not even the credentialed bastions of law, medicine, or academia are safe: AI can sift through oceans of content and serve up bespoke advice or coursework at a fraction of today's cost—and with quality that's closing in fast.
There are major questions about how much more powerful AI tools might become—and how soon. Anthropic's Dario Amodei and OpenAI's Sam Altman claim artificial general intelligence (AGI) could be only a year or two away. Meta's Yann LeCun is more skeptical, arguing that current models lack grounded physical understanding, durable memory, coherent reasoning, and strategic foresight, and Apple just published new research claiming that today's models perform only within the limits of their training data. Yet even if progress stopped tomorrow, the disruption is already underway.
To navigate this new landscape, leaders need to understand—and plan for—how automation will affect their businesses. That requires understanding which tasks and responsibilities are most likely to come under pressure and charting a course to move the enterprise up the intelligence value chain before time runs out.
What Is Not at Risk of Automation?
Academic researchers and practitioners have extensively debated which jobs and tasks are most vulnerable to automation. Some threats are obvious: self-driving vehicles may soon be in a position to displace millions of ride-sharing, bus, and truck drivers. Meanwhile, language translation, swaths of creative writing, design, and even everyday coding are being handed off to AI.
In February, Anthropic shared revealing user stats: although the chat format naturally steers people toward human augmentation, about 43% of interactions already represented some form of automation, in which users ask the AI to perform a task directly as opposed to helping them iterate and think it through. That share will keep climbing as modular AI agents enter the workforce, trading data and coordinating tasks through protocols like MCP. Environments that are extensively measured or codified—whether through laws, tax codes, compliance protocols, or streams of sensor data—face the greatest near-term risk of being handed over to machines.
AI research pioneers Ajay Agrawal, Joshua Gans, and Avi Goldfarb argued in 2018 that as AI advances, the last bastion of human advantage will be judgment—the ability to weigh options and make decisions under uncertainty. Yet that insight hands us an impossible homework assignment: pinning down exactly what qualifies as judgment at any given moment.
Tasks that demand human judgment today—choosing a medical treatment, reviewing a legal contract, scripting a film that nails the zeitgeist—could soon pass to AI as models tap richer data and greater compute. Nor can we assume people will always prefer a human therapist, counselor, or mediator, according to recent research. An AI counterpart can operate around the clock, at a fraction of the cost, and—aside from a handful of human superstars—may offer more consistent quality.
So, how can we separate the tasks AI will automate next from those that will require new breakthroughs in AI technology to do so? To answer that, we must go back to first principles and revisit where it all began.
From Lab Contest to Industrial Revolution
Back in the mid-2000s, computer scientist Fei-Fei Li saw that the field of computer vision, which is focused on enabling computers to 'see' and interpret images, was dealing with a bottleneck: algorithms were pixel-starved, ingesting too little visual data to reach human performance. Her solution was refreshingly brute-force: she built ImageNet—a vast, meticulously labeled image trove assembled with help from Amazon Mechanical Turk. But her true stroke of genius came in 2010, when she bolted a global leaderboard onto the dataset—transforming image recognition into a gladiatorial contest for researchers.
For two years, the annual leaderboard inched forward.
Then, in 2012, Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton blew the competition away. Using two off-the-shelf NVIDIA GTX 580 graphics cards, the trio from Toronto was able to train a breakthrough convolutional neural network in just a few days—a groundbreaking approach that proved you could bend computer-vision history on a grad student budget.
That moment ended the decades-long AI winter, put neural nets at the center of progress, and revealed the playbook the field still runs on. First, gather relevant data—roughly 14 million labeled images in the ImageNet case. Next, rely on metrics to quantify and drive progress. Last, flood a model with data and GPU muscle until it teaches itself, a formula that has carried AI from categorizing objects to writing fluent prose and, most recently, to reasoning, planning, and wielding external tools in today's emerging 'thinking' systems.
Data, Reward, Compute
The framework that propelled the image recognition breakthrough is far more general than most realize. It can be invoked whenever we can a) define the task environment and assemble its data—be it a corpus of text, a repository of images and video, logged driving miles, or streams from a robot's sensors; b) specify a target reward, explicit ('did the model predict the next word?') or implicit (inferred from observing human behavior); and c) provide the computational power to let the system iterate.
Stack those three ingredients and you get a general-purpose automation engine. Two data trends now accelerate the flywheel. First, models can mint limitless synthetic examples—for instance, generating virtual 'driving miles' that cover every oddball scenario, rather that relying on data from real world drivers. And second, AI is increasingly fielded across a variety of devices and sensors—on phones, in cars, and elsewhere—as a low-cost surveyor, capturing and quantifying real-world signals that were once too expensive or impractical to measure.
If you can shoehorn a phenomenon into numbers, AI will learn it and reproduce it back at scale—and the tech keeps slashing the cost of that conversion, so measurement gets cheaper, faster, and quietly woven into everything we touch. More things become countable, the circle resets, and the model comes back for seconds. That means that any job that can be measured can, in theory, be automated.
Measurement Too Cheap to Meter
Economist Zvi Griliches's landmark 1957 study of hybrid corn adoption gives us a sharp lens on what comes next. Farmers first planted the pricey seed only on their best acres—where the yield jump easily covered the extra cost and learning curve of using a new product. As hybrids improved and word spread, even thin-margin fields soon cleared the benefit-cost bar. With AI, the investment into measuring things follows the same payoff curve. When turning reality into data is expensive, companies tend to only invest in the headline cases—credit-card fraud, algorithmic market-making, jet-engine prognostics.
But AI now slashes the cost of precise measurement, making continuous, fine-grained sensing the default. Lightweight models run beside the sensors, trimming bandwidth and latency, while synthetic data fills gaps when the real world is slow or awkward to capture. Each extra decimal place quickly pays for itself: tiny error cuts multiplied across millions of AI-driven decisions add up fast. As precise measurement gets cheaper, ever-slimmer benefit streams pencil out, and tasks once too minor to monitor slide into the automation net.
Not only may we soon have intelligence too cheap to meter, we'll also be measuring ever more of the world to expand—and continuously upgrade—what that intelligence can reach. We already live in the era of 'artificial-metrics intelligence,' where anything we can quantify is swiftly queued for automation.
Thriving Despite Unknown Unknowns
Humans are evolutionary generalists, selected to navigate half-drawn maps. We don't merely survive unknown unknowns—we thrive on them, and that resilience is our defining edge. Over countless generations we fine-tuned our vocal cords and social brains until language emerged—opening the door to cumulative knowledge, abstract reasoning, and symbolic thought. From there we pushed beyond our biological limits, forging tools that stretched our senses, expanded our memory, and multiplied our abilities.
But the cornerstone of our advantage is our highly plastic, densely wired prefrontal cortex. This neural command center lets us spin endless 'what-ifs,' rehearse counterfactual futures, and pivot strategy the instant conditions shift. Short of a true singularity, even quantum machines will struggle to match our talent for open-ended, cross-domain counterfactual planning.
As AI accelerates progress, it creates fresh unknown unknowns, so our maps keep being redrawn. Meanwhile, it routinizes the predictable—much as mechanized farming lifted us from subsistence—freeing more of our counterfactual brainpower for higher-level problems.
AI will also struggle in domains where measurement verges on the impossible—witness the decade-long, globe-spanning effort the Event Horizon Telescope needed to capture a single black-hole image, and the still-unsolved challenges of probing extreme-scale physics, Earth's deep mantle and abyssal oceans, or live cellular interactions inside the human brain. It will also lag where measurement is throttled by privacy, ethics, or regulation; where society requires transparent reasoning—at least until model interpretability catches up; and where people simply prefer a human touch. Yet, as with hybrid corn adoption, future generations will keep revisiting the cost-benefit calculus for each of these—and may reach conclusions very different from ours.
But one crucial carve-out in what can be measured may prove decisive: tasks that defy quantification because their outcome odds are fundamentally unknowable—the realm of Knightian uncertainty, where you can't assign any probabilities because the risks themselves are undefined. Scaling a startup, allocating capital or talent into highly uncertain ventures, containing a novel pathogen, setting central bank policy during a financial regime shift, drafting AI ethics, inventing a new artistic medium, igniting a fashion trend, or creating a new genre-bending blockbuster—all sit in zones where probabilities vanish. Some creative acts and discoveries amount to little more than clever recombinations of the familiar, but the truly ambitious hinge on our singular ability to envision genuinely new and complex counterfactual worlds.
The list is fluid—tasks drop off the moment they become measurable, and new ones surface just as quickly. Each shift forces painful economic and social adjustments, squeezing more work into a superstar economy that concentrates outsized rewards at the peaks of creativity, talent, and capital. Yet AI offers a paradoxical gift: by democratizing education and serving as everyone's personal copilot, it hands more people than ever the tools to reach those peaks. Jobs themselves will keep evolving, and any breakthrough that turns the unknown into the countable will scale and be imitated at meme speed.
For leaders steering their organizations through this turbulent transition, what lies beyond the spreadsheet? It's everything that won't fit in a cell: the skills that refuse to be tallied, the open-ended problems with no reliable precedent, the intangibles—trust, taste, and the subtle dimensions of quality and experience—and the conviction to press ahead even when every metric says 'wait.' Manage only what you can measure, and you surrender the most valuable ground to rivals who cultivate what can't be counted. Amar Bose, the sound and electrical engineer who founded the Bose Corporation, proved the point: while others worshipped spec-sheet numbers, he zeroed in on how music sounded to people in real rooms—a quality no existing metric could catch—and in doing so, he rewrote the rules of the audio industry.
Directionally, the prescription is simple. Back wildcard bets with fuzzy ROI, reward teams that reframe problems and lean into the unknown, and rotate talent through roles that confront uncertainty across R&D, new markets, and complex customer, partner, and policy interactions. Carve out slack time and engineer cross-team collisions to spark serendipity and idea recombination. Treat those pockets of planned ambiguity not as liabilities, but as strategic assets.
Only leaders who pay attention to what is measurable—and, more crucially, to what stubbornly isn't—will be ready when the next shift arrives.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
20 minutes ago
- Yahoo
SoftBank founder proposes $1trn AI and robotics hub in US
SoftBank Group founder Masayoshi Son is pursuing a plan to establish a $1trn industrial complex in Arizona, the US, to create a major hub for AI and robotics manufacturing, reported Bloomberg. The Japanese investment firm has approached Taiwan Semiconductor Manufacturing (TSMC) to play a key role in the project, though the specifics of TSMC's potential involvement remain unclear. The proposed complex, dubbed 'Project Crystal Land,' envisions a high-tech manufacturing hub akin to China's Shenzhen, individuals familiar with the plan told the publication. The park could include production lines for AI-powered industrial robots, they said, though the discussions remain private. SoftBank has also engaged with Samsung Electronics executives to gauge interest, alongside a range of other technology companies, the report added. SoftBank officials have held discussions with US federal and state government officials, including US Secretary of Commerce Howard Lutnick, to explore potential tax incentives for companies investing in the industrial park. Son has compiled a list of SoftBank Vision Fund portfolio companies, such as robotics firm Agile Robots SE, that could establish production facilities at the site. The project's feasibility depends on support from the Trump administration and state authorities. While the envisioned cost could reach $1trn, as previously reported by the Nikkei, the actual scale will hinge on participation from major technology firms, Bloomberg's report said. If successful, Son has suggested the possibility of developing additional advanced industrial parks across the US. TSMC, which has already begun mass production at its first Arizona factory as part of a $165bn US investment, indicated that SoftBank's project does not impact its existing plans in Phoenix. Representatives from SoftBank, TSMC, and Samsung declined to comment, and the US Commerce Department did not immediately respond to Bloomberg's inquiries. In March 2025, SoftBank agreed to acquire Ampere Computing, a US-based semiconductor design company, for $6.5bn. As per the deal, Ampere Computing will be acquired by SBG through its subsidiary Silver Bands 6 (US) Corp. Ampere Computing is engaged in developing high-performance, energy-efficient processors tailored for cloud computing and AI workloads. The company currently employs approximately 1000 engineers. "SoftBank founder proposes $1trn AI and robotics hub in US" was originally created and published by Verdict, a GlobalData owned brand. The information on this site has been included in good faith for general informational purposes only. It is not intended to amount to advice on which you should rely, and we give no representation, warranty or guarantee, whether express or implied as to its accuracy or completeness. You must obtain professional or specialist advice before taking, or refraining from, any action on the basis of the content on our site. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Yahoo
24 minutes ago
- Yahoo
How to ask for a raise: 6 mistakes that can hurt your chances — and what to do instead
Only 13% of full-time employees requested a raise in 2023, but 66% of those who did received one, according to a May 2024 Federal Reserve survey. It goes to prove the old saying right: Ask, and you shall receive. Still, asking for a raise at work can be a stressful conversation for many employees. If you ask at the wrong time or in the wrong way, you could hurt your chances. From asking too soon to giving ultimatums, even minor missteps can work against you. In many ways, it's a skill of its own. Here are six mistakes to avoid, along with practical tips to approach salary discussions with confidence and increase your odds of success. Before asking for a raise, it's essential to establish a strong relationship with your manager. Otherwise, the conversation won't go far. 'Before giving a raise, I would want to know what my employee is doing and have a regular cadence where we're meeting consistently,' says Patrice Williams-Lindo, a career coach and manager of five. 'If it's the first time we're meeting, it's impractical to ask for a raise then and there. It's like asking someone to marry you on the first date,' she explains. Set up regular one-on-one meetings with your manager if you haven't already done so. Use that time to keep them updated on your progress, share your wins and make sure you're aligned on goals, including earning a salary increase. It will also give your manager a clear understanding of how you and your work contribute to the team. Do you work for a great organization? Nominate it as one of America's Top Workplaces. Even after you've established a relationship with your manager, avoid requesting a raise in an arbitrary one-on-one meeting, via email or during a casual conversation. Your manager should know the conversation is coming. 'It shouldn't be an ambush,' Williams-Lindo says. 'When you book the meeting, say, 'I'd like to talk about compensation and share the results I've driven,'' she advises. Schedule a dedicated meeting and clearly state its purpose. That way, your manager has time to prepare and come to the conversation with the right mindset. Timing can significantly impact whether your raise request is granted. If you're unsure when your company typically handles raises or promotions, bring it up during your one-on-ones, then use the intervening time to work toward that raise. 'If I'm setting myself up for mid-year, then I need to start at the beginning of the year — if not before — building that case,' Williams-Lindo says. Map out what you want to achieve for the year ahead and start gathering proof points early, so by the time you discuss a raise or promotion with your manager, your case will already be well established. Avoid asking for a raise out of fear or personal financial pressure. Instead, keep the focus on your performance and value. Williams-Lindo suggested saying something like, ''It's been X months — here's what I've done, the caliber of my work and the outcomes I've delivered. That's why I believe I'm qualified for X, Y or Z.' That gives off a different energy than, 'Hey, I don't know how I'm going to pay my rent, and I need a raise.'' Center the conversation on your professional achievements and value. Know what the market is paying for your role and what peers in similar roles (even at other organizations) are making. If you can demonstrate the gap, you're more likely to have a productive outcome. Giving an ultimatum when asking for a raise is a major mistake. Even if you do get the raise, this type of communication can antagonize your manager, damage your professional reputation and ultimately undermine your value. 'This might sound like, 'I need a raise, or I'm quitting.' I appreciate that bravado, but for me, it's a bit of a turn-off, because how did we get here?' Williams-Lindo says. Rehearse your talking points. Calm, professional conversations are far more effective than explosive ones. Be open to negotiation and prepared to compromise on the final amount. Even if you've thoroughly prepared and presented your case well, your raise request could still be denied. If you're told "not right now," ask your manager for specific feedback. There may be outside factors influencing the decision or specific goals you'll need to meet to be considered for a raise in the future. A good manager will explain the reason — whether it's because the company is facing a downturn or because there are performance gaps to work on — and then help you to fix them. Set up a follow-up time — in three to six months — to revisit the conversation. This will indicate that you are engaged and goal-oriented. If your manager is dismissive and doesn't provide clear, achievable paths to advancement, it may be time to reassess if this is still the right place for your growth. This article originally appeared on USA TODAY: how-to-ask-for-raise
Android Authority
26 minutes ago
- Android Authority
Can we trust Google Maps to get us anywhere anymore?
Andy Walker / Android Authority 🗣️ This is an open thread. We want to hear from you! Share your thoughts in the comments and vote in the poll below — your take might be featured in a future roundup. This past weekend, I was driving home from a wonderful break in the country. I usually use Google Maps to guide me home on longer drives, even if I have driven that route before. It's more of a comfort than a necessity. However, it tried its level best to take me on an unwanted adventure. To get home from the tiny town I visited, I had a choice between a faster toll route or the longer scenic mountain pass that runs alongside it. I selected the 'Avoid tolls' toggle before I left, thinking that the app would suggest I use the pass. After all, it was the best alternative available. That wasn't the case at all. Maps instead suggested two hugely divergent routes that would've taken me two hours north or 90 minutes south of the best route. Absentmindedly, I came so close to taking the suggested deviations. And if I hadn't curiously glanced at my Android Auto display, I would've ended up further from home. This incident opened my eyes to the larger problem affecting Google Maps' navigation reliability, especially on the open road. Having read countless takes online about Maps guiding drivers on obtuse routes, I'd say it's a fairly wide-reaching issue. So, if you have a story to tell, I'd love to hear it. I'm sure other readers would too. Here are the questions: Has Google Maps ever gotten you lost? If so, tell us about the experience. What's the worst place Google Maps has ever taken you? Do you feel that Maps has grown more inaccurate and less reliable over time? What is your preferred navigation app on Android, and why do you pick it? Would you pay a premium for a Google Maps that offers smarter navigation and more contextually aware features? Be sure to vote in the poll below, too! Has Google Maps ever gotten you lost while driving? 0 votes Yes. Google Maps consistently gets me lost. NaN % Yes. Once or twice, but not very often. NaN % No, Google Maps has always been a reliable guide for me. NaN % I don't use Google Maps while driving. NaN % 👇Sound off in the comments with your Google Maps adventure anecdotes and opinions.
