Personal

From spell check to agentic AI

I once hated communication because I have dyslexia. Every tool on this timeline removed a barrier between how I think and how others need to receive it. Agentic AI didn't just remove the last barrier — it turned the disability into the advantage.

The Shift

Assistive technology for the dyslexic designer

Every dyslexic professional has a secret stack of tools. The spell checker you can't live without. The voice assistant that lets you draft emails by talking. The AI that turns your fragmented thoughts into polished prose.

We don't talk about it because it feels like cheating. It's not. It's assistive technology — the same category as glasses, wheelchairs, and hearing aids. Tools that let you do what your brain already knows how to do, through a medium that doesn't fight you.

Here's my timeline. Forty years of technology that slowly, then suddenly, turned a communication disability into a communication superpower.

1985

Spell Check

Microsoft Word ships with spell check. For the first time, a dyslexic writer gets real-time feedback on spelling errors. Limited — can't handle phonetic misspellings like 'shud' for 'should'. But a start.

Could write documents without asking someone to proofread every sentence.

1995

Autocorrect

Word's autocorrect starts fixing common mistakes automatically. 'teh' becomes 'the' without thinking. Small but transformative — reduces the cognitive load of catching your own errors while trying to think.

Writing speed increased. Less shame in draft documents.

2004

Text-to-Speech (Read Back)

Screen readers and TTS tools like NaturalReader become mainstream. A dyslexic designer can hear their own writing read back — catching errors the eye misses. Proofreading by ear, not by eye.

Self-editing became possible. Less reliance on others to check work.

2007

iPhone & Predictive Text

The iPhone puts autocorrect in your pocket. Predictive text suggests words as you type. For a dyslexic thinker, the phone starts finishing your thoughts — guessing what you mean from fragments.

Text messages and emails stopped being anxiety-inducing.

2011

Siri & Voice Assistants

Speech-to-text goes mainstream. Dictate instead of type. For someone whose thoughts flow faster than their spelling can keep up, voice input is liberation. Think it, say it, done.

Bypassed the spelling bottleneck entirely for quick communication.

2012

Grammarly

Context-aware grammar and spelling correction. Unlike basic spell check, Grammarly understands what you meant — even from phonetic misspellings. Catches tone, clarity, and structure. A writing partner, not just a red underline.

Professional writing became possible without a human editor. Career-changing.

2016

Dyslexia-Specific Fonts & Tools

OpenDyslexic font, Ghotit spell checker (built by dyslexics), coloured overlays in browsers. Tools designed by people who understand the problem, not just the symptoms.

Reading code and documentation became less exhausting.

2020

Figma & Visual-First Design

Design tools that let you think spatially, not textually. Drag, arrange, connect. A dyslexic thinker's natural mode — Material reasoning made digital. Design without writing a spec first.

Could express ideas visually and have others extract the spec.

2021

GitHub Copilot

AI code completion. Write a comment describing what you want, and the code appears. For a dyslexic developer, this bridges the gap between knowing the logic and remembering the syntax.

Coding speed caught up to thinking speed. Syntax stopped being a barrier.

2023

ChatGPT / Claude

Full conversational AI. Describe what you want in plain language — messy, fragmented, dyslexic language — and get structured, polished output. The AI translates how you think into how others need to read.

Documentation, proposals, emails — all became one conversation away from done.

2024

Claude Code & Agentic Workflows

AI agents that read your context files, understand your design system, and execute code. CLAUDE.md becomes the narrative layer — encode your intent once, agents act on it repeatedly. Communication at scale.

Entire websites built through conversation. This site is the proof.

2025

Claude Voice Mode

Speak to Claude, it writes. Not dictation — conversation. Hold spacebar, describe what you want in fragmented, dyslexic, verbal thinking, release. The agent translates it into structured output. The input channel shifts from typing to talking. For someone whose thoughts flow faster than their fingers can spell, voice is liberation.

The keyboard becomes optional. The voice becomes the interface. I→N→C through speech.

2025

MCP, Figma MCP, Design Tokens as AI Infrastructure

Design systems become machine-readable infrastructure. Tokens, components, and context files that AI agents consume directly. The design system communicates on your behalf — to humans AND machines.

One person, four AI agents, production output. The team is encoded, not hired.

2026

Agentic AI: The Full Stack

Custom AI agents (UI Designer, UX Designer, FED Developer, UX Researcher) that execute design and development under your direction. A Stakeholder Simulator that practices your pitch. Voice cloning that lets your avatar speak for you. The dyslexic thinker doesn't need to write — they need to direct.

Communication is no longer the bottleneck. Vision is. And that was always the strength.

The Parallel Shift

Text is dying. Visual language is rising.

While assistive tech was helping dyslexic thinkers survive in a text-based world, something bigger was happening: the world itself was shifting to visual.Every decade, another layer of text got replaced by a visual interface. We're not adapting to a text world anymore. The world is adapting to us.

We are creating a new language — a visual language for visual thinkers, future thinkers. And the people who were “disabled” in the old text medium are native speakers of the new one.

Text World (dying)
Visual World (rising)
1984
Command Line Interface
Macintosh GUI

Text commands replaced by icons, windows, mouse. Visual thinkers could finally use computers without memorising syntax.

1990
Written manuals
Multimedia CD-ROMs

Learning through video, animation, and interaction instead of reading. The first hint that text wasn't the only medium.

1995
Text-heavy web (HTML)
Flash & rich media

The web goes visual. Animation, interactivity, spatial layouts. Designers start building experiences, not documents.

2007
Desktop software menus
iPhone touch UI

Direct manipulation replaces menu hierarchies. Pinch, swipe, tap. The interface becomes physical, spatial, intuitive. No reading required.

2010
Written specifications
Wireframes & prototypes

Design communication shifts from text documents to visual artefacts. Show, don't tell. The spec is the prototype.

2013
Email & documents
Slack, Emoji, GIFs

Even business communication goes visual. Reactions replace replies. A thumbs-up says more than a paragraph.

2016
Static mockups
Figma real-time collaboration

Design becomes a shared visual space. Multiple people thinking visually together in real-time. No document handoff.

2020
User stories (text)
FigJam, Miro, visual boards

Product thinking goes spatial. Sticky notes, flow diagrams, visual mapping. The boardroom becomes a canvas.

2023
Text prompts
AI image generation

Describe a vision in words, see it instantly. Text becomes the input, visuals become the output. The ratio flips.

2024
Code as text files
Visual development (V0, Stitch)

Describe a UI, get a UI. Code is generated from visual intent. The developer's medium shifts from text to vision.

2025
Design tokens as JSON
Live theme switching, visual preview

Design systems become visual infrastructure. Swap tokens, see the result instantly. The system communicates visually.

2026
Documentation
AI agents that see and build

Agents consume visual context, generate visual output. The architect directs visually. Text becomes metadata, not the medium.

2030+
Screens and windows
AR/XR spatial interfaces

The interface IS the world. Spatial computing, gesture control, ambient UI. Text becomes a relic. Visual-spatial thinkers — dyslexic thinkers — are native speakers of this new language.

The Convergence

Two timelines, one conclusion

The assistive tech timeline shows tools helping dyslexic thinkers survive in a text world. The visual language timeline shows that text world disappearing.

They converge in 2026. Agentic AI is both the ultimate assistive technology AND the tool that builds the visual future. An architect who thinks in spatial relationships, connections, and narrative — who was “disabled” in the text era — is perfectly equipped for the visual era.

AR/XR, spatial computing, gesture interfaces, ambient UI — these are all visual-spatial mediums. The designers who will build them think in exactly the way dyslexic brains are wired: Material reasoning, Interconnected reasoning, Dynamic reasoning. Not sequential text processing.

Yesenia Perez-Cruz at IDS 2026 described the next generation of design systems as a language: Grammar (design system rules) + Vocabulary (domain objects) = Sentences (composed UI). AI assembles the sentences. The interface isn't written — it's spoken into existence. The text-based specification becomes a visual-spatial composition. The grammar is structural. The vocabulary is contextual. The output is visual.

For dyslexic thinkers who always struggled with text grammar but excelled at spatial composition, pattern recognition, and narrative structure — this new language isn't learned. It's native.

Dyslexia wasn't a disability. It was early adoption of a language the world hadn't invented yet.

The Pattern

Each tool removed a barrier. AI removed the last one.

Spell check fixed individual words. Grammarly fixed sentences. Voice input bypassed typing. ChatGPT fixed paragraphs. Claude Code fixed entire workflows.

The pattern is clear: each generation of assistive technology moved one layer up — from characters to words to sentences to documents to systems. Agentic AI operates at the systems level.It doesn't fix your spelling. It consumes your intent and produces the entire output.

For a dyslexic thinker, this is the inflection point. The barrier was never the thinking — it was the translation from thought to medium. When the medium is a conversation with an AI agent, the translation cost drops to near zero. What's left is pure cognitive contribution: vision, connections, narrative, possibilities.

The MIND strengths — Material, Interconnected, Narrative, Dynamic reasoning — were always there. The tools finally caught up.

For Others Like Me

If you're a dyslexic designer or developer

You're not cheating by using AI. You're using assistive technology. The same way a short-sighted person uses glasses, you use tools that bridge the gap between how you think and how the world needs to receive it.

Your ability to see patterns across domains, to think in spatial relationships, to simulate scenarios dynamically, to encode complex ideas as narrative — these are cognitive strengths that the design industry desperately needs. The spelling and sequential processing are what the tools handle.

Build your stack. Grammarly for writing. Claude Voice Mode for drafts. Claude Code for implementation. Custom agents for your team. CLAUDE.md for your context. Design tokens for your system. Encode your narrative. Let the agents execute.

You were never bad at communication. You were using the wrong medium. AI is the right one.

I once hated communication because I have dyslexia. Agentic AI is the assistive technology that lets me enable others to do what I was never able to. I don't design. I don't code. I encode the narrative they need to act without me.
Lincoln Mitchell — the N in I<N>C

The tools caught up. Now what?

If you're a dyslexic designer, developer, or architect navigating AI — or if you're building tools for neurodivergent thinkers — I'd love to hear from you.

/