Demultiplexing Ability: How AI Can Align Function, Design, and Independence

There’s a strange tension around artificial intelligence right now.

On one side, it’s hailed as transformative.

On the other, it’s criticized for replacing jobs, flattening craft, or amplifying noise. The debate swings between hype and fear, as if AI must either revolutionize everything or quietly undermine it.

What gets lost in that argument is a simpler question: what should AI actually be for?

In my world — adaptive design, aging in place, functional independence — the answer is clear. AI should not replace expertise; it should extend it. It should not erase human judgment; it should make that judgment more accessible. It should not displace care systems; it should reduce the friction inside them.

That conviction didn’t start with AI. It started with consolidation.

Before the Ability Curve was ever about guidance, it was about proof. If you design adaptive products around individual diagnoses, you fracture the opportunity into small, disconnected markets — stroke here, arthritis there, Parkinson’s in another category, aging somewhere else entirely. Each appears niche. Each appears limited.

But that’s not how functional reality works.

Grip weakness is grip weakness.

Fatigue is fatigue. Instability is instability.

The functional drivers overlap across diagnoses and across age. If you consolidate symptoms instead of labels — if you count everyone who struggles with hand strength, everyone who compensates for reduced stability, everyone who adapts around pain — the market stops looking small. It reveals itself as massive.

The Ability Curve was created out of that need for consolidation. Not just philosophically, but economically. Bring those overlapping functional drivers together and you can see the full opportunity. You can design around the true size of the population that shares those constraints — and justify investing in better geometry, better force transfer, better control. You can design a knife around hand challenges broadly, not around a diagnosis narrowly.

Years ago, when I worked in telecom marketing with Tellabs in the Dense Wave Division Multiplexing group, we dealt with a similar principle. DWDM takes multiple light frequencies — each carrying separate streams of data — and consolidates them into a single fiber for efficient transmission. The efficiency comes from consolidation: you maximize capacity by combining the streams. But when the signal reaches its destination, you demultiplex it. You separate the frequencies back out so each stream can be interpreted precisely.

That was the same mental model.

First, consolidate the functional data. Arthritis, limb difference, MS, age-related decline, neuropathy — bring them together at the level of grip demand and control. Prove the scale. Prove the overlap. Design for that consolidated signal.

Then demultiplex it.

Because real people don’t live as a consolidated market. They live as individuals. They don’t experience “the hand impairment sector.” They experience layered constraints — arthritis plus vision decline, fatigue plus reduced grip, stroke recovery plus balance concerns, age-related weakness combined with endurance loss.

The Ability Curve allowed consolidation. AI allows demultiplexing.

Once you understand the full signal, you can reverse-engineer it for the individual.

Imagine being able to describe a task in plain language: “I can’t chop carrots without pain.” Not a diagnosis. Not a clinical code. Just a lived challenge. An intelligent system doesn’t need to know the ICD-10 classification. It can deconstruct the task — how much force is required, what kind of motion, how much stabilization, where fatigue accumulates, what the risk exposure is.

It becomes conversational demultiplexing. You describe the friction. The system helps isolate which strand of the consolidated signal applies to you.

Take it further. Drop in a recipe — not because you need cooking instructions, but because you need sequencing guidance. What should be cut first? What can be stabilized before heat is introduced? What steps reduce repetition? How do you stage the task so fatigue doesn’t spike halfway through? For someone already managing pain or limited grip, that kind of structured clarity isn’t a convenience. It’s the difference between making the meal and skipping it.

This isn’t about automating cooking. It’s about narrowing the functional load to match the person.

And this is where AI becomes genuinely interesting — not as spectacle, but as a thought partner.

Most of the ecosystem supporting people with functional limitations is chronically understaffed. Nonprofits are stretched thin. State agencies manage overwhelming caseloads. Clinicians are constrained by reimbursement models that compress interaction time. Expecting each of these professionals to deeply understand every adaptive product on the market is unrealistic — not a failure of commitment, but a failure of system design.

Manufacturers face a related friction. They design for consolidated functional drivers, but must communicate in demultiplexed language — tailored to specific users, specific funding pathways, specific clinical conversations. Without a bridge, the burden of translation falls back on the individual: the person already navigating limitation.

That’s backwards.

If AI can serve any ethical purpose here, it is this: to act as the dynamic demultiplexer. To take a consolidated understanding of ability and break it down in real time for a person’s specific context — not to tell someone what to buy, not to promise outcomes, but to help them articulate their situation in functional terms. To help them understand how a tool aligns with their constraints. To help them walk into a physician’s office and say, “This improves my ability to safely prepare meals because it reduces the grip demand and repetitive sawing motion that aggravates my fatigue.”

That’s not marketing. That’s clarity.

The broader vision isn’t about a single knife or a single platform. It’s about reducing the fragmentation in a system that desperately needs coherence. Consolidate where scale and design require it. Demultiplex where precision and dignity require it.

In an ecosystem where AI could help match functional challenges to adaptive solutions — across manufacturers, nonprofits, and agencies — the value wouldn’t accrue only to companies. It would accrue to case managers who need a fast way to narrow options, to nonprofits who need scalable guidance without hiring ten more staff members, to users who need structured thinking without drowning in search results. AI is criticized for replacing jobs. In under-resourced sectors, it can do the opposite. It can create capacity. It can free professionals to focus on high-value human judgment while the system handles structured narrowing and task decomposition.

That, to me, is what AI should be for.

Not spectacle. Not noise. Not synthetic content at scale.

Precision in service of independence.

The Ability Curve consolidated the signal. Using AI to demultiplex it — conversationally, task by task, layer by layer — is the next step. Not to remove the human from the equation, but to give the human better leverage.

Because independence isn’t built on labels.

It’s built on alignment

---

If this message resonates, please consider trying my NULU Navigator BETA and provide some feedback on the experience.