Angehefteter Tweet
Built Different
12.5K posts

Built Different
@BuiltDiffly
Where discipline meets mindset. | Sport. Self-dev. Sharp thinking.
Beigetreten Haziran 2023
44 Folgt193.6K Follower
Built Different retweetet

Embedded AI (agentic operations, agentic process automation, whatever we're calling it this month) is basically business process automation with better brains.
For decades, enterprises ran critical repeatable processes on rules engines and RPA. Take a process that has to run itself reliably and make it work. Now we call it 'agentic.' New name, same ambition.
The difference is real capability though. Old BPA could follow rules. Embedded AI can reason through exceptions, learn from past runs, pull meaning out of unstructured data. Not just a rebrand.
But the hard problems are the same ones BPA teams were solving 15 years ago. Governance, observability, human-in-the-loop checkpoints, managing failures at scale. We just have to solve them again with systems that are way more powerful and way less predictable.
The teams shipping fastest right now are usually the ones who learned from that era. They don't treat governance as an afterthought because they already got burned once.
I keep saying the agentic industry renames the same things over and over. This is one of those cases, but at least this time the capability underneath actually justifies the new name.

English
Built Different retweetet

Mining will create more millionaires than AI over the next decade. Yet everyone is chasing chips and software. Here's why the real money is in what powers all of it:
The world is obsessed with AI. Every portfolio, every headline, every allocation decision points in the same direction.
But here's what most investors are missing.
AI runs on infrastructure. Infrastructure runs on metals. Copper, nickel, palladium, platinum. The physical materials that make electrification, data centres, and energy grids possible.
Supply is constrained. From discovery to production, the average mine takes around 18 years. And demand is growing faster than at any point in the last 30 years.
This is the setup for a commodity supercycle. And historically, supercycles are where generational wealth gets built in the resource sector.
Power Metallic sits at the centre of this thesis. A zoned polymetallic system with copper, nickel, and PGEs. Only a handful of comparable systems exist on the planet. Well funded with C$50 million raised and 100,000 metres of drilling through 2026.
While everyone fights over the next AI stock, the real asymmetric opportunity might be sitting underground in Quebec.
$PNPN $PNPNF
English
Built Different retweetet
Built Different retweetet
Built Different retweetet
Built Different retweetet
Built Different retweetet
Built Different retweetet
Built Different retweetet
Built Different retweetet

Everyone thinks we're running out of copper and nickel. We're not.
Remember when the world was sure we were running out of oil? We weren't. We'd just run out of the technology to find it efficiently.
Critical minerals are in the same place right now. The earth has plenty. What's constrained is our ability to find it, permit it, and bring it to market fast enough.
Here's the twist: AI is actually solving its own supply problem. Real breakthroughs in mineral exploration and processing. Recycling is scaling. Copper and nickel can both be recycled.
This isn't "finite resource, permanent crisis." It's massive demand hitting a temporary supply bottleneck.
The bottleneck isn't geology. It's capital, timing, and willingness to fund exploration.
That's the trade. Not scarcity panic. Timing arbitrage.
Full conversation: youtube.com/watch?v=m2mb7S…
$PNPN $PNPNF

YouTube
English
Built Different retweetet
Built Different retweetet
Built Different retweetet
Built Different retweetet
Built Different retweetet
Built Different retweetet
Built Different retweetet

Claude Fable vs Claude Opus 4.8 (side by side)
Same prompt, but big delta in output + token use + time spent to getting there.
See for yourself + "spend the day as the president"
Imo Fable for high-quality one shots vs Opus for highly-specified tasks/interactive sessions
Fable vs Opus
> Time: 27 min vs 5 min
> Tokens: 134k vs 64k
> Cost: ~$6.70 vs ~$1.60
> LOC: 1.5k vs 990
> Errors: 0 vs 0
> Self-QA: Fable playtested its own game in a headless browser. Opus caught + fixed its own crash bug.
English
Built Different retweetet
Built Different retweetet

I've appeared before the Texas Supreme Court 14 times.
I've represented hundreds of institutional property owners across Texas - school districts, hospital systems, universities, apartment communities, commercial real estate.
If there is one thing I have learned in all of that work, it is that the most expensive problems are the ones nobody thought to ask about.
If I could sit across from every school superintendent, hospital CEO, and university facilities director in the state and give them three things to investigate before accepting their next construction project, these would be the three:

English
Built Different retweetet






















