Maybe the solution isn’t reforming the big machine. Maybe it’s replacing it entirely. Thoughts on productivity, funding, a fast car, Ironwood, and more.
Productivity is the most socially acceptable form of procrastination.
We’ve built entire industries around “getting things done” instead of doing what matters.
Most people aren’t productive—they’re just busy avoiding hard decisions. They spend hours tweaking dashboards, fiddling with calendars, optimizing workflows…
Not out of discipline. Out of fear.
Fear of committing. Fear of being wrong. Fear of actually doing the work.
Real builders don’t optimize. They execute.
Clean systems don’t build great companies. Clarity does.
3 Things Worth Sharing
1. Funding
There’s a dangerous myth in early-stage entrepreneurship:
“If I raise more, I’ll win.”
In reality, venture capital doesn’t solve your problems—it amplifies them.
Bad hiring? You’ll do more of it.
No product-market fit? You’ll scale dysfunction.
No clear value prop? You’ll spend a year building features that nobody needs.
VC’s aren’t looking for perfect models, they’re looking for momentum, clarity, and asymmetric upside. The smartest ones are backing systems and founders, not slide decks and hype.
Here’s what I’ve learned (and seen) over two decades building and investing:
The best rounds don’t chase headlines—they chase heat. Warm intros and hype don’t beat real traction and founder clarity.
A lean, focused team with a clear niche will outperform a bloated, funded team almost every time. Especially in markets that punish burn.
VC’s don’t fund ideas. They fund inevitability. Your job as a founder is to make your execution feel inevitable. Not optional. Not probable. Inevitable.
If you can’t make money without VC, you probably won’t with it.
2. Speed
The McMurtry Spéirling, a fan-assisted electric prototype, has set a new benchmark by becoming the fastest car to lap the Top Gear test track, achieving a time of 55.9 seconds, surpassing the previous record set by a Renualt F1 Car by 3.1 seconds.
Powertrain: Dual electric motors delivering approximately 1,000 bhp
Weight: Under 1,000 kg
0–60 mph: Approximately 1.4 seconds
Top Speed: Electronically limited to around 150 mph
Downforce: 2,000 kg generated by a fan system, even at a standstill
Noise Level: Fan system emits about 120 dB, comparable to a jet engine
This achievement underscores the potential of electric vehicles in high-performance applications, challenging conventional automotive engineering and pushing the boundaries of speed and design.
3. Big Shift
Google just dropped Ironwood—its most powerful TPU yet, built specifically for AI inference at planetary scale. For most people, this goes over their heads.
Key Specifications:
Compute Power: Each Ironwood chip delivers up to 4,614 TFLOPs of FP8 performance.
Memory: Equipped with 192 GB of High Bandwidth Memory (HBM) per chip, offering 7.2 TB/s bandwidth.
Scalability: Supports configurations up to 9,216 chips per pod, achieving a total of 42.5 exaFLOPs.
Efficiency: Delivers 2x the performance per watt compared to its predecessor, Trillium.
SparseCore: Includes an enhanced SparseCore accelerator for processing ultra-large embeddings, beneficial for recommendation systems and scientific computations.
Training gets all the headlines, but inference is where the real game is played. Ironwood is built not to impress, it's built to serve AI to billions in real time. This marks a significant leap in AI hardware, emphasizing real-time decision-making capabilities.
The race isn’t to build the biggest model. It’s to have the fastest deployment.
“Startups, not states, will define the next generation of defense.”
Shyam breaks down why legacy defense contractors can’t keep up. Procurement is now a liability. Speed is deterrence. Software is the new battlefield.
Takeaways:
Sovereignty depends on velocity.
Defense must be deployable in hours, not years.
Startups like Palantir, Anduril, and Shield AI are the new prime contractors
This is what defense looks like when it's built like a startup. In an age when 51 becomes 5, times are changing, and we’re heading back to a multi-input defense approach. The big players are frustrated, but we’re in this mess because of them.
Brandon Tseng, ex-Navy SEAL and co-founder of Shield AI, breaks down autonomous warfare: drones that fly without GPS, make decisions on the edge, and win through real-time intelligence.
“We don’t need more humans in the loop. We need faster loops.”
AI Radar
Suno v3 – Studio-quality songs with vocals and lyrics in seconds.
FireCut – AI cuts long-form videos into viral-ready shorts.
Reka AI – New contender in frontier models, with strong performance and clean architecture deployable anywhere.
Limitless AI – Most wearable Ai assistant for meetings, notes, and context tracking.
Question I’m Pondering
What if governance should be built like startups. Fast iterations, small teams, and local ownership?
Maybe the solution isn’t reforming the big machine. Maybe it’s replacing it entirely.
Until Next Time
Back to building.
Your challenge:
Look at one system you rely on—your schedule, your news, your workflows, and ask: Is this designed for clarity or control?
See what happens when you rebuild it from scratch.
See you next Thursday. — Jacques
P.S. If you enjoyed this, share it with someone who thinks like you.
Check out other things I’m busy with:
We Edit Podcasts – A full-service B2B podcasting agency that helps brands turn listeners into customers.
Hunting Portal – Your next hunting adventure starts here. Outfitters, taxidermists, gear heads, and more.
Halalisana — We help township-based businesses grow, scale, and thrive in South Africa. We offer conferencing, co-working, and a coding academy.
Melck Group – Buy undervalued real estate in a small coastal town in South Africa.
