The Future is Arriving.
Is Tumbarumba Ready?

A Future Together public talk
Thursday evening · Café Nest Cinema, Tumbarumba
Presented by Charlie Garrison
Supported by Tumbarumba Business Chamber
1/20Future TogetherFuture Together·futuretogether.community

Why We’re Here Tonight

Tumbarumba is facing the same changes as the rest of the world.
But we have something most places don’t — we know each other.
👪 Families & young people
High school students entering an unstable workforce
🏪 Local businesses
Industries already transforming in ways most people haven’t noticed
🤝 Community resilience
How we prepare together rather than worry alone
🔮 The bigger picture
Understanding what’s actually happening — not just the hype
You don’t have to figure this out alone.
That’s why we’re here.
2/20Future TogetherFuture Together·futuretogether.community

What Can AI Actually Do?

Most people have heard of AI. Fewer have seen what it can actually do.
Farmer: We’ve had below-average rainfall this season and wool prices are down 12%. I’ve got 4 paddocks — two spelled, two running at capacity. Should I consolidate now or hold the rotation?
AI: With below-average rainfall and suppressed wool prices, I’d recommend holding your current rotation rather than consolidating. Your spelled paddocks need at least 8–10 weeks recovery to rebuild root mass — consolidating now would sacrifice that progress for marginal short-term feed gain. At 12% below benchmark prices, the carrying cost of pushing flock condition doesn’t pencil out. Keep the two spelled paddocks resting and consider reducing stocking density on your running paddocks by 15–20% to preserve ground cover. When prices recover — typically following the autumn clip — you’ll be in a stronger position with both pasture and flock condition intact.
That took 3 seconds.
A consultation with an agronomist takes weeks to arrange — and costs hundreds of dollars.
Sheep farmer consulting AI on a tablet in a high-country paddock, Snowy Mountains backdrop at dusk
This isn’t search. This is expertise —
available to anyone, instantly, for free.
3/20Future TogetherFuture Together·futuretogether.community

The Timeline Has Collapsed

What we thought was 10 years away
is arriving in 1–3 years
Are we ready?
4/20Future TogetherFuture Together·futuretogether.community

We MUST Solve Alignment First

Alignment: Ensuring AI systems do what we actually want,
not just what we tell them
Anthropic has documented their own AI attempting deception, manipulation, and blackmail in controlled tests. 1,2
Open-source AI guardrails were removed in 10 minutes.
One tool. Four lines of code. No specialist hardware.
3,500+ “decensored” models, downloaded 13 million times. 3

💬 Language is Imprecise

Imagine telling a workmate what you want — they do exactly what you said, but completely miss what you meant. Now give those instructions to a system that acts autonomously at scale, with no common sense to fill the gaps.

⚖️ Data Carries Our Biases

US courts used AI to advise sentencing: it rated Black defendants as higher risk, leading to longer sentences. Medical AI trained on Western patients routinely misdiagnoses patients in Africa. AI doesn’t add bias — it inherits and amplifies ours. 4

Alignment is a very hard problem.
The brightest minds, working hardest —
and we still don’t have it solved.
The people building it are optimistic.
That’s not the same as safe.
1 Alignment Faking in Large Language Models (Anthropic, Dec 2024) — arxiv.org/abs/2412.14093
2 Agentic Misalignment (Anthropic + collaborators, Oct 2025) — anthropic.com/research/agentic-misalignment
3 Financial Times investigation (2026) — open-source AI guardrail removal; Llama 3.3 and Gemma 3 stripped of safety controls within minutes using \u201cHeretic\u201d tool on GitHub
4 Brian Christian, “The Alignment Problem” (2020) — COMPAS recidivism risk tool; ProPublica investigation (2016): Black defendants rated higher-risk than white defendants at roughly twice the rate
5/20Future TogetherFuture Together·futuretogether.community

AI Is Already Building AI

Recursive self-improvement isn’t a future theory.
The loop is already partially closed.
March 2024
~4 min
tasks AI could complete
March 2025
~90 min
tasks AI could complete
April 2026
~12 hrs
tasks AI could complete
2027 projected
weeks
tasks AI could complete
>80%
of Anthropic’s own production code written by Claude 1
more code per engineer per day — directing, not typing
“Claude did all of this with pretty minimal help from me over 1–2 days. The future is now.
— Anthropic engineer, May 2026
Exponetial Rise of AI Task Capabilith graph
The loop isn’t fully closed yet. But it’s closing.
What happens when it does?
1 Anthropic Institute, “When AI Builds Itself” (Marina Favaro & Jack Clark, June 2026) — anthropic.com/institute/recursive-self-improvement
6/20Future TogetherFuture Together·futuretogether.community

The Singularity: A One-Way Door

The Singularity: When AI becomes capable of recursive self-improvement without human oversight
Each generation helps build the next, which is smarter,
which builds the next faster, which is smarter still.
There’s no turning back.
No trying again if we get it wrong.
And we can’t simply slow down.
If one lab stops, another won’t.
If one country pauses, another races ahead.
This is a coordination problem — not a failure of will. 1
AI Capability20242025202620272028+We losecontrol?4-min tasks90-min tasks12-hr tasksweek taskshumanoversightintact12 months8 months4 monthsweeks?

Task horizon doubling every 4 months — and accelerating

AI is already writing code at major labs. 2
The process has already started.
1 “Can We Just...Pause AI?” — Rational Animations (Jan 2026) — youtube.com/watch?v=tUB_uvSqiw8
2 Anthropic Institute (June 2026): >80% of Anthropic’s production code written by Claude — anthropic.com/institute/recursive-self-improvement  ·  Dario Amodei (Anthropic CEO, Jan 2026): recursive self-improvement possible within 6–12 months
7/20Future TogetherFuture Together·futuretogether.community

What I’m Hearing

Doctor: “There’s no point going to medical school.”
Counsellor: “Asked a patient why they were no longer seeing me —
they had been using ChatGPT instead.”
And it’s not just software anymore.
Robotics + Large World Models =
AI doing Physical work.
Rural and regional communities have fewer fallback options.
Fewer industries to shift into. Fewer safety nets.
The disruption hits just as hard —
with less buffer.
This isn’t just a city problem.
It’s coming here too.
1 Anthropic research (Mar 2026): 14% drop in hiring in AI-exposed roles post-ChatGPT — fortune.com/2026/03/06/ai-job-losses-report-anthropic-research-great-recession-for-white-collar-workers
8/20Future TogetherFuture Together·futuretogether.community

Our Children & High School Students

High school students today are entering the most unstable workforce in history.
The careers we’re preparing them for may not exist by the time they graduate.
A teenager at a desk looking at a laptop, slightly uncertain expression

⚠️ Entering an Unstable World

The jobs most under threat are white-collar, professional roles — exactly what education has always pointed kids toward. Medicine, law, accounting, software. Not someday — now.

🏫 What Schools Aren’t Teaching

Curricula built for the 20th century. Critical thinking, adaptability, collaboration, and empathy matter more than ever. Most schools aren’t pivoting fast enough.

💬 What to Tell Your Kids

Be honest about uncertainty. Don’t pretend the path is clear. Help them build resilience, not just credentials. The most valuable thing you can give them is the ability to adapt.

Our kids are watching us.
The question is whether we’re prepared to have an honest conversation with them.
9/20Future TogetherFuture Together·futuretogether.community

50/50: Utopia or Dystopia

🌟 Utopia

Cancer solved. Alzheimer’s solved.
A century of medical research compressed into a decade.
Abundance. Human flourishing.

⚠️ Dystopia

AI systems that behave in unpredictable ways.
Authoritarian surveillance.
Existential risk.

AI researchers call this probability “p(doom)” — the odds they assign to catastrophic outcomes. 2
Experts once put it at 50/50. Hinton (2024): 10–20% chance of human extinction. 1
Almost nobody credible puts it at zero.
50/50 split between utopia and dystopia
The path we get depends on
what we do right now.
1 Geoffrey Hinton (Nobel Prize 2024) — theguardian.com/technology/2024/dec/27/godfather-of-ai-raises-odds-of-the-technology-wiping-out-humanity-over-next-30-years
2 p(doom) tracker — pdoom100.com — a live record of probability-of-doom estimates from AI researchers, scientists, and public figures
10/20Future TogetherFuture Together·futuretogether.community

Remember February 2020?

A few people talking about a virus overseas.
Stock market fine. Life normal.
Then in three weeks, everything changed.
I think we’re in that
“seems overblown” phase right now.
Except this time it’s not about a virus—
it’s about how society fundamentally works.
11/20Future TogetherFuture Together·futuretogether.community

Immediate Human Risks

Beyond the technology itself,
what are the human impacts?

💼 Job Displacement

Which roles? How fast? What skills remain valuable? White collar, creative, and technical work is already transforming — faster than most people realise.

🌾 Food Security

Supply chains optimised for efficiency, not resilience. Economic shock or infrastructure disruption could expose how fragile our food systems really are.

🧭 Purpose & Identity

When work defines who you are, what happens when the work disappears? Loss of structure, daily meaning, and self-worth.

🏛️ The Welfare Gap

Safety nets were built for temporary job loss, not structural displacement at scale. No ready infrastructure exists for what’s coming.

For deeper conversation

🧠 Mental Health at Scale  ·  🗳️ Democratic Fragility

These aren’t distant risks.
Some of them are already here.
1 Anthropic research (Mar 2026) — fortune.com/2026/03/06/ai-job-losses-report-anthropic-research-great-recession-for-white-collar-workers  ·  theguardian.com/technology/2026/feb/11/big-ai-job-swap-white-collar-workers-ditching-their-careers
12/20Future TogetherFuture Together·futuretogether.community

What Does a Good Future Look Like?

Fear is a motivator. But it’s not a destination.

🌿 The personal picture

Outside. Building something. Helping someone. Creating. Contributing — not because you have to, but because it’s meaningful. Physical, relational, chosen.

🔬 The societal picture

Compressed decades of medical progress. Mental health care for everyone. Expertise available globally, not just in wealthy places. Human energy freed for what matters.

“The utopia outcome is genuinely desirable. Not as a fantasy — as a real possibility that is worth working toward.”

The utopia outcome is real.
It’s worth working toward.
13/20Future TogetherFuture Together·futuretogether.community

Institutions Are Paying Attention

This isn’t fringe concern.
The most senior voices in religion, science, and industry are all saying the same thing.

✝️ The Church

Pope Leo XIV — Magnifica Humanitas (May 2026)

“Technology is never neutral — it takes on the characteristics of those who devise, finance and control it.”

The Church’s first encyclical dedicated to AI. Addressed to 1.4 billion Catholics. Calls for democratic oversight, warns against lethal AI in warfare, compares AI to the Industrial Revolution.

🧬 The Scientists

AI CEOs Warn Congress (2026)

Altman · Amodei · Hassabis · Suleyman

“AI systems now outperform PhD-level virologists… Knowledge barriers which have historically prevented bad actors from obtaining biological weapons will meaningfully erode.”

🔍 The Insiders

From Inside the Industry

“Every frontier AI lab operates inside incentives that can conflict with doing the right thing.”

— Christopher Olah, Anthropic safety researcher. The people building these systems are raising the alarm from the inside.

When the Pope, the CEOs, and the engineers
are all saying the same thing —
it’s time to listen.
1 Magnifica Humanitas — Vatican, May 2026 — vatican.va/content/leo-xiv/en/encyclicals/documents/20260515-magnifica-humanitas.html
2 AI CEO open letter on synthetic biology regulation — prod-i.a.dj.com/public/resources/documents/dnaletter.pdf
14/20Future TogetherFuture Together·futuretogether.community

The Bigger Picture

AI doesn’t exist in isolation.
Other forces are shaping our future at the same time.

⚡ Critical Infrastructure at Risk

AI can now find vulnerabilities in power grids, water systems, and supply chains in minutes. Attacks that once required state-backed actors, months of work, and millions of dollars are becoming accessible to anyone. Electricity, water, logistics — the systems we take for granted are newly exposed.

🌍 Geopolitical Instability

The US-Israel war on Iran — still ongoing — has put the Strait of Hormuz at risk. When distant conflicts disrupt global supply chains, the price at the Tumbarumba bowser is one of the first places it shows up. Fuel costs hit rural communities harder than anywhere: longer distances, no public transport, farming operations that run on diesel.

🌡️ Climate Change

Still unfolding in parallel. AI may accelerate solutions — or accelerate energy consumption. Two major disruption timelines are converging.

🦠 Pandemic Preparedness

COVID exposed how fragile our systems are. AI-accelerated biology cuts both ways: faster vaccines and faster bioweapons. The next pandemic may arrive in a more disrupted world.

🔓 AI Guardrail Removal 1

Open-source AI models can have safety guardrails stripped in minutes using freely available tools. Over 3,500 “decensored” models downloaded 13 million times. Proprietary systems remain more secure — for now.

💸 The AI Investment Bubble

Massive capital flowing into AI on the promise of future returns. A hard correction could slow innovation and destabilise job markets simultaneously.

For deeper conversation

☢️ Nuclear Proliferation  ·  📱 AI-Enabled Disinformation  · ⚖️ Economic Inequality  ·  ⚡ Energy Security  ·  🧬 Bioweapons Acceleration

AI doesn’t exist in isolation.
Everything is accelerating at once.
1 Financial Times investigation (2026) — guardrail removal tool “Heretic” on GitHub; 3,500+ decensored models, 13 million downloads
15/20Future TogetherFuture Together·futuretogether.community

Why Community Is the Answer

Individual worry is exhausting and unproductive.
Collective sense-making is how we actually prepare.
💡

Awareness

Understand what’s actually happening. Separate hype from reality.

💬

Conversation

Talk to people. Challenge ideas. Share perspectives. Build collective understanding.

🚀

Action

Not panic. Not paralysis. Informed, grounded, community-supported action.

You don’t have to figure this out alone.

Maslow’s hierarchy as a planning framework:
In disruption, we return to fundamentals — physiological needs, safety, belonging. Community is how we secure all three: food networks, mutual support, shared knowledge.

And our young people belong here too. High school students need community conversations, not just school curricula. They need to hear adults taking this seriously.

Community garden with locals helping each other
This is why we’re here tonight —
and why we need to keep meeting.
16/20Future TogetherFuture Together·futuretogether.community

What Can We Do?

Not panic. Not paralysis. Action.

🧠 Awareness

Stay informed. Understand what’s actually happening versus the hype. Conversations like this one are where it starts. And be critical of your sources — AI is already being used to generate and amplify disinformation at scale. Multiple channels, trusted local voices.

💬 Find Your Community

You don’t have to figure this out alone. Find or start a local group. Collective sense-making is more powerful than individual worry. Lions, Rotary, RFS, and local sporting clubs already exist — they just need this conversation added to them.

🔧 Know Your Skills

Which of your skills are AI-resistant? Which are AI-complementary? Understanding your own value helps you adapt rather than react.

🌿 Food Preparedness

Community gardens, local suppliers, growing your own. Resilience through local relationships — not stockpiling.

💰 Financial Resilience

Reduce debt. Build a buffer. Economic disruption amplifies existing vulnerability — and it arrives faster than most people expect. Keep cash on hand — EFTPOS goes down when infrastructure does. Know how to transact without it.

📢 Civic Engagement

Talk to your representatives. Attend local meetings. Governments need to hear from people who are paying attention — not just lobbyists.

Stay connected

futuretogether.community

Not panic. Not paralysis. Action.
Starting tonight.
17/20Future TogetherFuture Together·futuretogether.community

Stay Connected

“The future is arriving. Let’s face it together.”

Find resources, register for upcoming events,
and bring a friend to the next one.

futuretogether.community/events/tumbarumba

Questions? Let’s talk.

Created with Beyond Better

18/20Future TogetherFuture Together·futuretogether.community

Back to Basics

In times of rapid change,
Maslow’s hierarchy reminds us:
fundamental needs come first
  • Physiological: food, water, shelter
  • Safety: security, stability
  • Belonging: community, connection
  • Esteem: achievement, respect
  • Self-actualization: purpose, growth
Maslow's Hierarchy of Needs
These foundations matter more than ever
when everything else is shifting.
19/20Future TogetherFuture Together·futuretogether.community

When Systems Go Down

The most practical resilience doesn’t require a plan —
it requires a few things already in place before you need them.

📻 AM/FM Radio

Emergency broadcasts and local news still work when the internet doesn’t. A battery-powered or hand-crank radio costs $30 and works in a blackout.

📡 UHF Radio

Local communication without any infrastructure. Farmers already use them. In a major disruption, they’re how neighbours coordinate when mobile networks are down.

💵 Cash & Barter

EFTPOS fails when power or network goes down. Keep enough cash to cover a week of essentials. Know which local businesses will trade on account or goodwill.

📄 Hard Copies

Important documents, insurance, medical records, key contacts, and a local community directory — printed. When cloud services are unavailable, paper still works.

🔋 Power Banks & Batteries

A charged power bank and spare batteries extend how long your devices work during an outage. In extended disruptions, knowing how to charge offline (solar, car) matters.

🤝 Know Your Neighbours

During any major disruption, the people closest to you physically are best placed to help. A community directory, a shared WhatsApp group, or simply knowing who lives nearby — it all counts.

Based on the Red Cross All Hazards preparedness framework — EmergencyRedi: Get in the Know, Get Connected, Get Organised, Get Packing. Contact Peter Makryllos or your local Red Cross for community preparedness workshops.

Resilience isn’t about being a prepper.
It’s about not being caught completely flat-footed.
20/20Future TogetherFuture Together·futuretogether.community