AI makes you FEEL faster while making you 19% slower
Your weekly dose of AI & startup news on our path to 1,000 Aussie startups.
It is 3:00 am.
You are the President of the United World, sitting in the Global Spire, watching your holographic dashboard glow.
Outside, the city is on fire.
You can hear the thumping. Thousands of them. The Kangaroos have breached the perimeter.
Your robotic sphincter tightens.
Your generals are calling. Your advisers are screaming. Your dashboard is insisting everything is fine.
This Week’s Line-Up
🔍 The Decision
🔥 Exceptional Events
🤖 Techies Weekly Update
📼 Weekly Video: Does AI really make us more productive?
👀 Claude? No. Get ready to get CLAWED
🦘Memes of the Week
Written by: Jun Kai Chang (Luc), Julia Ponder, Dr Sam Donegan, Shivang Shekhar
📂 The Decision
INBOX: 📩 1 NEW MESSAGE
FROM: Department of Efficiency (War Room)
SUBJECT: The breach report
At 2:17 am, Perimeter Control stopped recognising kangaroos as “hostile”.
The turrets did not shut down. The dashboard stayed green. The system was “online”.
It just stopped firing at the right thing.
Why?
A junior cadet merged an AI-written patch straight to main.
The patch was titled: “Optimise Roo Detection Logic v3”.
It was beautiful. Clean. Commented. Confident.
It was also wrong.
Normally, a senior would catch it. Tonight, the seniors were buried in a mountain of other AI-generated changes. Review queue, on fire. Context switching, feral. Everyone exhausted.
So the patch shipped.
And now the Kangaroos are doing unspeakable things to the South Lawn.
Your strategy seemed sound, lets use AI to fight back against the vile beasts! The AI will make us more productive!
But that was the ‘F$#k around’ phase. Now we’re in the ‘Find out’ phase.
And we’re finding out.
Does AI make us more productive?
The answer is: yes, sometimes. Also no, sometimes. We’ve learned. It depends what you asked it to do and whether you checked its work.
📂 FILE 1: Where AI actually helps
Drafting and routine knowledge work
In controlled studies, people using ChatGPT for professional writing tasks finished faster and produced higher quality work.
In real workplaces, a large study of customer support agents found an AI assistant increased issues resolved per hour on average, with the biggest gains for less experienced workers.
What that means for the war effort:
AI is excellent at first drafts, summaries, comms, and repeatable responses.
AI is a productivity boost when the work is easy to check quickly.
If we had limited AI to those jobs, you would be asleep right now.
📂 FILE 2: Where AI bites you in the bippy
Complex, high-context engineering
A recent randomised study on experienced open-source developers working on real issues in their own repositories found that allowing early-2025 AI tools made them slower, about 19% longer to finish tasks. Developers still felt faster. Reality disagreed.
Why this matters for us:
Perimeter Control is not a blank project. It is a complicated system with years of history, brittle edges, and the kind of logic where “mostly correct” is the same as “we are about to be eaten”.
AI can generate plausible code fast.
But in complex systems, the expensive part is not typing. The expensive part is verifying.
That is where we failed.
File 3: Why the dashboard stayed green while we lost the lawn
You may have heard someone shouting that productivity is up about 5% since the AI era began.
There has been a real jump in human productivity in some quarters. For example, BLS reported nonfarm business labour productivity up 4.9% in Q3 2025 at an annual rate.
The DORA report found that higher AI adoption was associated with an estimated decrease in delivery throughput and stability. In plain terms, teams can look faster while shipping gets shakier.
And GitClear’s analysis of 211 million changed lines of code reported trends like less refactoring and more copy/pasted code in recent years, which is the kind of thing that quietly turns your codebase into a swamp.
So yes, the dashboard can be green.
It can still be lying about whether we are winning the war.
🛠️ The actual lesson of the Kangaroo breach
The red phone on your desk screams. You pick it up. It’s General M1AI, calling from the Front Lines. You can hear techno music and laser blasts in the background.
“Hello President. AI did not betray us.
We did the classic human thing where we found a powerful tool and immediately used it in the one place where mistakes cost lives.
We used AI to make us faster at producing changes, but we did not upgrade the part of the system that makes changes safe:
review capacity
testing
verification
and the courage to say “no, this is not ready”
So the juniors shipped faster.
The seniors became overwhelmed.
And one unverified AI patch turned our perimeter into an automatic door.”
🗳️ YOUR ORDERS?
The clock is ticking. The Kangaroos are pounding on the blast doors. The army is waiting for your command.
[PUSH TO SIGN YOUR DECREE]
🔴 Option A: The “Ban” Protocol.
Ban AI for junior cadets.
You get fewer rookie disasters in the short term.
You also throw away real productivity wins in drafting-heavy work, and you teach the next generation nothing about using the tool safely.🔵 Option B: The “Cyborg” Protocol.
Allow AI, but change the rules of war.
AI can draft. Humans must verify.
No AI code merges without tests that prove it.
Keep changes small so review does not become a landfill.
Measure outcomes, not output. Count Kangaroos stopped, not patches shipped.
🟢 Option C: The “Luddite” Protocol. “Return to Tradition.” We go back to 2021. Manual weapons only. We will be slower, but goddammit we’ll be human.
🔍 Did you make the right choice?
The Kangaroos do not care how many lines of code we generated.
They care whether the turret fires.
Choose.
🏥 MedHack is running again soon!
Last year’s MedHack was pure chaos (the good kind): teams shipped real prototypes in a weekend and they were genuinely impressive.
This year we’re doing it again with MEDHACK: FRONTIERS (6th edition) and it’s the hackathon you’ll still be talking about next year. Team Formation Night is Sun 15 Feb (5–9pm), then the Hackathon Weekend is Sat 21 – Sun 22 Feb. If you want to build something that matters, meet people you actually want to work with, and pitch to a packed room of health, tech and venture leaders, grab a ticket soon. Spots are limited by design. 👉 Early Bird tickets here.
🔥 Exceptional Events (OMG there’s too many)
1. Melbourne | AI Builder Co-working x StartSpace
🗓 Saturday 7th Feb | ⏰ 10:00am | 📍 StartSpace
This event is great for those who want to work on their AI products, see what others are building, get to know Melbourne’s awesome AI community more, and just hang out and have a great time!
2. Cursor Cafe
🗓 Saturday 14th Feb | ⏰ 9:30am (Choose a slot) | 📍 Bottari
We’re taking over a local café (Bottari) and inviting the Melbourne community to come build with us. Bring your laptop, your ideas, and come spend some time collaborating, debugging, or just chatting with other Cursor users.
3. Melbourne | AI Builder Co-working x Stone & Chalk
🗓 Saturday 21st Feb | ⏰ 9:30am | 📍 Stone & Chalk
This event is great for those who want to work on their AI products, see what others are building, get to know Melbourne’s awesome AI community more, and just hang out and have a great time!
4. MYMI x MLAI: MedHack Frontiers
🗓 21st–22nd February 2026
Australia’s most chaotic health-tech hackathon. Team up with Hackers, Hustlers, Hipsters & Healers to solve real medical challenges and push digital health beyond buzzwords.
👉 Early Bird tickets here.
5. Fayl Tales Live: What Almost Broke Me
🗓 27th Feb 2026 | ⏰ 5:30 PM | 📍 Stone & Chalk Melbourne
Most startup events are about how people made it. This one is about how they almost didn’t... This is an intimate live event for founders, operators, and builders who are curious about what really happens behind the highlight reel.
6. buildDay: Build & Deploy a Working AI App in 6 Hours
🗓 28th Feb 2026 | ⏰ 8:30 AM | 📍 Stone & Chalk Melbourne
Build a real AI-powered web app in one day. Deploy it live. Take it home. Never built an app before? Perfect. That’s exactly who this is for.
7. MedHack: Frontiers Pitch Night
🗓 4th March 2026 | ⏰ 6:00 PM | 📍 Monash College City Campus
After one intense weekend of clinicians, engineers, designers, and students speed-running healthcare problems, we’re closing MedHack: Frontiers the only way that makes sense: put the prototypes on stage and make the teams defend them in public (politely, mostly).
8. Vibe Marketing: Hack Your Way To Page #1 on Search Engines
🗓 7th March 2026 | ⏰ 10:30 AM | 📍 Stone & Chalk Melbourne
Get your business on page #1 on Seach Engines and ChatGPT while you sleep basically. Come spend a Saturday with us and learn a fast, practical system that works. This workshop is built for founders, business owners & marketers. No fluff. No jargon. Just a clear playbook and code you can use the same day.
👉 Only 12 spots available each month
🛠️ TOOL SPOTLIGHT: Ertas AI
What it is: The “Canva” for building custom AI models.
Right now, training your own AI model is like using Photoshop in the 90s (complex, technical, and expensive). Ertas AI fixes this. It is a user-friendly platform that lets enterprises train, fine-tune, and deploy private models in-house.
No more API fees: Stop paying per token.
No more data leaks: Your data never leaves your server.
No PhD required: Technical and non-technical teams can finally collaborate on model behavior.
🤖 AI Bits for Techies
This week: AI Productivity Reality
AI is moving SO FAST! So we moved the heavy lifting to a dedicated weekly report. If you want the deep dive, including weekly research paper summaries, new open-source tools, and model architecture updates, this is for you. This week find out whether AI really boosted our productivity.
View the Techies Report: Click Here
📼 Weekly Video: : Does AI really make us more productive?
Video: The AI productivity paradox
Length: 6.14mins
About: It breaks down the gap between the 40 percent efficiency gains we see in isolated tasks – like a dev smashing out a boilerplate – and the actual economic output of a whole company.
🚀 Claude? No. Get ready to get CLAWED
While everyone else is waiting on a waitlist, Peter Steinberger just open-sourced the future. Next Thursday, we’re deploying a deep dive into OpenClaw. It’s the most aggressive implementation of the digital personal assistant pattern we've seen yet. The "Claw" is coming to your local machine → are you ready to compile?
Subscribe so you don’t miss the learning. Got another case? Drop it in #general in slack (or email us at hi@mlai.au if the link does not work) and collect Roo Points.
🦘 AI & Startup Memes
Share your memes in #random-and-memes and if they’re funny we’ll give you free stuff.
By Vaughan Shanks:
Some funny post of what AI thinks of their boss:
By Ty Coon:
By Jun Kai Chang (Luc):
