Amjad Masad, CEO of Replit, just made a provocative argument: the economy doesn't need true AGI (artificial general intelligence). Instead, he argues for "functional AGI"—AI systems that are good enough at enough tasks to be economically transformative, even if they're not truly intelligent in the way humans are.

This might sound like semantic hairsplitting, but it's not. It's a fundamental reframing of what businesses should expect from AI and when they should expect it. And if Masad is right, a lot of companies are waiting for the wrong thing.

For years, the AI industry has been chasing AGI—a system that can do anything a human can do, across any domain, with human-level or superhuman performance. OpenAI's stated mission is to build AGI. Google DeepMind is working toward it. Anthropic talks about it cautiously. It's the North Star.

The promise is seductive: build AGI once, and it solves everything. It's the universal employee, the perfect consultant, the tireless analyst. It drives your trucks, writes your code, manages your supply chain, and answers your customer emails. All you have to do is wait for someone to crack it.

Masad's argument is simpler and more pragmatic: we don't need to wait for that. We already have AI that's good enough at enough things to change how businesses operate. It's not perfect. It makes mistakes. It can't do everything. But it doesn't need to.

He calls this "functional AGI"—AI that functions well enough across a broad range of tasks to be economically useful, even if it's nowhere near human-level general intelligence.

If you're a business owner waiting for "real AI" to arrive before you commit to a strategy, this reframe should change your thinking.

True AGI might be five years away. It might be twenty. It might require breakthroughs we haven't made yet. Betting your business strategy on its arrival is betting on a timeline nobody can predict.

Functional AGI is already here.

GPT-4 can draft emails, summarize documents, write code, analyze data, and generate content. It's not perfect at any of these tasks. But it's good enough that millions of people use it daily to get work done faster. That's functional AGI in action.

Claude can review contracts, suggest edits, help with research, and assist with strategic thinking. It won't replace your lawyer, but it'll help your lawyer work faster. That's not full intelligence—it's functional usefulness.

For businesses, the question isn't "when will AGI arrive?" It's "what can we do with the AI that exists today?"

Masad's framing aligns with what successful companies are already doing: they're not waiting for perfect AI. They're deploying imperfect AI where it delivers value.

You don't need an AI that understands everything. You need an AI that understands your customer emails well enough to draft responses your team can review and send in 30 seconds instead of writing from scratch in 5 minutes.

You don't need an AI that writes perfect code. You need an AI that writes code good enough that your developers spend less time typing boilerplate and more time solving hard problems.

You don't need an AI that makes perfect hiring decisions. You need an AI that screens resumes well enough to save your HR team from reading 500 applications and lets them focus on the 50 worth interviewing.

This is the 80/20 rule applied to AI: get 80% of the value from imperfect systems, rather than waiting for 100% from systems that don't exist yet.

Functional AGI has limits, and understanding those limits is critical.

It's inconsistent. AI performs brilliantly on some tasks and fails spectacularly on others, often unpredictably. You can't just deploy it and walk away—you need humans supervising the output, especially for high-stakes decisions.

It's narrow within breadth. AI today can do many things, but it can't do any of them with true human-level judgment. It can write a marketing email, but it can't tell you if the marketing email aligns with your brand strategy or if this is the right week to send it.

It's a tool, not a replacement. Functional AGI augments your team. It doesn't replace their judgment, creativity, or strategic thinking. The companies winning with AI aren't eliminating roles—they're redefining them.

If Masad is right—and the evidence suggests he is—then the entire framing of "AI strategy" changes.

Stop asking: "When will AI be smart enough to replace this role?"

Start asking: "What parts of this role can AI handle right now, and how do we reorganize work around that?"

Stop asking: "Should we wait for better AI before we invest?"

Start asking: "What value can we extract from imperfect AI today, and how do we scale that as the technology improves?"

The businesses that thrive in the next five years won't be the ones who bet on AGI arriving on schedule. They'll be the ones who ruthlessly exploit functional AGI while everyone else is waiting.

Here's what changes if you adopt the "functional AGI" mindset:

Lower the bar for "good enough." If you're waiting for AI to be 100% accurate before you use it, you'll never start. If you're willing to use AI that's 80% accurate with human review, you can start today.

Focus on volume tasks. Functional AGI shines where the task is repetitive, the stakes are moderate, and speed matters. Customer support. Data entry. Content drafts. Document review. These aren't glamorous, but they're expensive at scale.

Build review workflows, not replacement workflows. The winning pattern is AI generates, human reviews. Not AI replaces. Companies that try to eliminate human judgment entirely get burned when the AI makes a mistake nobody caught.

Experiment fast and cheap. If you're not expecting perfection, you can test AI in low-risk areas without massive investment. Try it on internal tools first. See what works. Scale the wins. Kill the failures.

The AGI vs. functional AGI debate isn't academic—it's strategic. If you believe we need true AGI before AI becomes transformative, you're going to wait years while your competitors move now.

Amjad Masad is arguing that the future is already here, it's just unevenly distributed. Businesses that recognize this will deploy imperfect AI systems today and iterate as the technology improves. Businesses that wait for perfect AI will find themselves behind companies that settled for "good enough" five years earlier.

True AGI might arrive eventually. But functional AGI is already delivering value to the businesses smart enough to use it. The question isn't whether AI is ready for your business. It's whether your business is ready for AI that's good enough, not perfect.

If you're still waiting for perfect, you're waiting too long.

_____________________________________________________________________________________________

Trying to figure out where AI fits in your business, or whether your current AI strategy is actually working? That's exactly what we do at Cue Crafted. We help small and mid-sized businesses cut through the hype and build AI strategies that deliver real value, not just expensive experiments. Schedule an intro today.