The Singularity Won’t Save Us (But It Might Help With Therapy)

Will AI expand access to care, or just replicate the systems that already fail us?

I Was a Teenage Cyborg

Last night I went to see futurist Ray Kurzweil speak at 92Y, in conversation with David Rose. It felt like time travel back to the baby-adult college version of myself who excitedly brain-ate my way through The Age of Spiritual Machines in a computer science lab. But I was really always more interested in the sociology of the internet (how are relationships formed in MOOs and MUDs?¹) than trying to build out programs in C++.

I’ve never really wanted to be a programmer.² But I’ve always been interested in learning about systems and how people behave inside them.

So, What’s the Singularity Again?

Kurzweil has created a bunch of important tech but seems best known for the singularity—the theoretical future moment when artificial intelligence surpasses human intelligence and begins improving itself recursively. The current forecast has this happening in 2045.

And AGI (artificial general intelligence) by 2029. AGI is AI that possesses the cognitive abilities and adaptability of humans. That’s just four years away. Which feels… unrealistically soon.

My Tiny Pencil Question (That Didn’t Get Picked)

I had a question I wrote on an index card with my Tiny Pencil but didn’t get asked:

If large language models still hallucinate, confidently generating false information, how do we get to AGI in four years?

I’m sure the answer has something to do with exponential growth on the abstract end and on the technical end, things like: feeding models better data, helping them access external sources to fact-check in real time, or giving them more of their own memory so they stop confidently reinventing reality every time they answer a question.

But, it still feels like hope/hype.

Not Replacement. Augmentation.

Kurzweil frames AI as augmentation, not replacement. We’re not being overtaken, we’re being enhanced. We don’t need Asimov’s Three Laws of Robotics to save us.

For example: our phones already augment our memories. I only know like 3 phone numbers by heart!

Eventually, he predicts, AI will become integrated with our biology: non-invasive interfaces that work like cognition extensions. Not Musk-y brain chips, but tools that act like externalized versions of executive function. Like:

• A system that boosts working memory, holding reminders, context, and next steps the way your prefrontal cortex tries but has a hard time doing under stress and overwhelm

• Attention-filtering glasses or earbuds that reduce sensory overwhelm by prioritizing relevant input

• Emotion-regulation tools that don’t just track your heart rate but detect when you’re freaking out and suggest something grounding before you’re fully consciously aware (when it might be too late)

Therapy Is Supposed to Help Rewire Your Brain, Right?

As I apply for Therapist School, I keep coming back to this: what if mental health support didn’t just happen in the brain, but alongside it? Post-it notes, Google Docs, AI conversations at 2 a.m. are already external cognitive tools. So maybe it’s not so dystopian to consider what the next step is.

Of course, if we leave it to Big Rich Tech Dudes with zero empathy and too much VC, “cognitive augmentation” will just become an expensive BrainProPlus subscription for wealthy people. But if we can manage to do it right, they could offer a new kind of access: not a cure, not a replacement, but scaffolding.

This isn’t just theoretical. There are real mental health tools already out there, like:

• Woebot, which uses natural language and CBT protocols, has shown reductions in depression and anxiety in clinical trials.

• FDA-cleared digital therapeutics³ like reSET (for substance use) and Somryst (for insomnia) deliver evidence-based treatment through software—not chatbots, but AI-driven behavioral platforms.

• Limbic, a clinical AI assistant, helps therapists triage patients, identify risk markers, and even write intake notes.

OMG IT’S GONNA ALTER YOUR DNA

Kurzweil also brought up the Moderna vaccine. It’s true: Moderna used machine learning to optimize its mRNA sequence in less than 48 hours after the COVID genome was released. The AI models helped simulate protein folding, select codons for human translation, and accelerate a design process that usually takes months.

That’s not the singularity. And it’s not what most people think of as AI. But it’s accelerating science, not replacing scientists.

Techno-Utopianism Hits a Wall (The Wall Is Made of Money)

We already live in a country where teachers have to DoorDash on weekends and therapists stop taking insurance because they can’t live on reimbursement delays. UBI would be amazing, but I doubt it will happen in the US. And while AI may create new jobs, no one ever says what those are. “Prompt engineer” isn’t a career path for most people. “Me but even more me” isn’t gonna pay my bills.

Kurzweil made an interesting point: a technology usually needs to be substantially better than what it replaces to be accepted. Self-driving cars may already be statistically safer than human drivers—but humans still don’t trust them. The PR narrative of “machine failure” is louder than the data no one reads. If a robot car hits someone once, it’s front-page news. If a human does, we barely notice.

And honestly, a lot of that is branding. If we called them “automated trams” instead of “driverless cars,” I don’t think people would be as freaked out.

So: AI is not going to save us. But it might help us survive in weirder ways. Especially if we focus on tools that augment human care vs replacing it.

---

Kurzweil: AGI by 2029.

Me: okay but it still hallucinates.

Also me:

but maybe machines hallucinating just means they’re closer to being human than we realized (whOaAaA……)

1. Sherry Turkle, Life on the Screen was a big one

2. OK but I got an A on my program that reduced deadlock in OS for my Operating Systems class. Still, I wanted it to be more about how they work on a UX/UI level than code.

3. I’ve worked on these, specifically! I wrote materials to help patients access this care and have insurance pay for it.