Exploring the Frontier of AI, Tech, and Innovation | We Stand on the Shoulders of Giants

Joined June 2021
College degrees won’t matter in the AI economy. Being a high-agency generalist will. Don’t specialize—build a toolkit for navigating complex systems. Markets reward those who steer AI swarms toward meaningful solutions. Your ancestors would trade places in a heartbeat.
1
13
60
10,222
AI Archive retweeted
The last discovery like ours was made about 18 years ago. So when we put out our grades, a lot of people simply didn't believe them. I get it. If you've never seen numbers like that from one system, a page of assays reads like a typo. So we stopped arguing. We moved the resource estimate up six months and had it done independently, to the 43-101 standard. Lion: about 4.75 million tonnes at 3.9% copper equivalent, over 85% Indicated. I told Red Cloud TV at Beaver Creek it was a great initial start. That's what it is. $PNPN.V | $PNPNF | $IVV1
1
13
2
33
2,503
AI Archive retweeted
We're using cosmic rays to look for copper. Muons rain down from space all the time and pass straight through rock. Dense rock slows them down a little. Put sensors down deep holes beside a deposit, count what gets through, and you get a density map of what's underground. Massive sulphide is heavy rock, so it stands out. We've deployed it at Lion. Henry Lazenby of The Northern Miner asked me how it works. Here's the short version. Why it matters: when the resource was built, some pockets near the Lion shoot weren't connected enough for the engineers to include. If Muon shows they are, we know exactly where to drill. --Terry $PNPN.V | $PNPNF | $IVV1
11
30
2,164
People I work with say I fail upwards. Ten years as an electrician. A first business that flopped. Then a job offer with no job description. Here's the whole path, failures included: At 21 I went to the UK for a gap year to be a personal trainer. I ended up a removalist. The company moved high profile clients, so I packed up homes for Madonna, Jude Law and Ozzy Osbourne, then drove the loads through eight countries at a time. Back home I was an electrician. In 2012, still on the tools, I started my first business: LED lighting imported from China. Two mates and I flew over to choose samples. We spent most of the money in Macau casinos on custom suits. Dumb and Dumber all over again. The tech was too early anyway. It was the first of a few businesses that didn't work. By 2015 I was ten years into the trade and wondering what came next. A close mate asked me a strange favour. Could I go and find some trees? I didn't ask questions. I found a few decent sized trees at a scrapyard and drove them to a childcare centre in South Melbourne. The owner wanted them on the top floor, so we tilted them on an angle, squeezed them into the lift, and built natural play for the kids. The owner was Mark Woodland. A while later I was doing electrical work at another of his centres. He told me he was starting a childcare software company and asked if I wanted a job. I asked what I'd be doing. He said he had no idea. But if I wanted a job, I had it. I said yes. I was his first hire. The company was acquired by a US private equity firm in 2020, and I stayed until 2022. Today my wife and I run three businesses together, and I'm a co-founder at Kismet with Mark, building in healthcare. The trees, the suits, the job with no description. Most of the good parts started with a yes. If I could tell my 10-year-old self one thing: You've got a crazy life ahead of you. Take every opportunity. Say yes more than you say no. Follow me @SammyRArmstrong for more on building something that matters.
1
10
11
1,344
AI Archive retweeted
Our own VP of Exploration refused to tell me our metallurgy results. I asked him what had come back. He said he could not tell me. I said what do you mean you cannot tell me. He said he needed confirmation first. He did not believe them. The numbers were far enough above what he expected that his instinct was not excitement, it was suspicion of his own data. So he sent the work back and had it run again. Three times in total. They held. SGS Canada reported recoveries of 98.9% copper, 96.8% platinum, 93.9% palladium, 88.9% silver and 85% gold. I have thought about that exchange a lot since. A geologist who will not repeat a good number until he has checked it three times is the most valuable person you can have on a project like this. Investors ask me why they should believe our grades. It is a fair question, and the honest answer is that you should not, not on my say so. Neither did we. That is the whole business in one story. You do not get to believe your own results. You test them until someone independent agrees, and then you publish what they found. $PNPN.V | $PNPNF | $IVV1
4
12
1
50
4,076
Your cold email replies are dropping, but what if your copy isn't the problem? The mailboxes sending your emails could be wearing out. This is one of the harder problems with running outbound at scale. A mailbox starts deteriorating. Bounces creep up. Fewer emails reach the inbox. But the campaign keeps running, so the team starts rewriting copy and testing new offers. Meanwhile, the mailbox that started it is still sending. @smartlead_ai's Auto-Replace is built to handle this before it becomes a bigger problem. You set the health rules once. By default, a mailbox gets flagged when its bounce rate goes above 4% over 14 days, or its reply rate drops below 0.3% across 500 sends. Mailboxes coming up on their one-year vendor limit get flagged too. Smartlead can then order a replacement automatically on a new domain at the same monthly rate, and warm it for 14 days before it goes near a campaign. Switch on auto-attach and it takes the old mailbox's place in every campaign. The old one stays live, so late replies still land. So sending capacity gets replaced before it costs you a reply. Because when you're managing dozens of inboxes, finding out something is broken shouldn't depend on someone noticing a drop in replies.
5
9
12
5,484
AI Archive retweeted
Denise Elgin wrote her number on a matchbook for a guy at a St Kilda nightclub. He assumed it was fake. It started with 555, and he thought those only existed in movies. He called anyway. 3 months later he proposed at a petrol station. Here is her story:
1
9
10
2,410
2 days ago, I hosted my first dinner in London, powered by @stanforcreators We brought together creators, operators, and founders with a combined audience of 900k+ and worked with household names like Microsoft, Meta, and LEGO. My takeaways: > London has an incredible talent pool, but bringing the right people together takes intention and proactivity > The more our feeds fill with AI-generated content, the more the pendulum swings harder towards IRL communities > You really can just shoot a DM to a stranger to invite them to a dinner and have a great time. Highly recommend Thanks, for coming! @czamaru @aryanagab @sashacayward @alstras @_danreynolds_ Julius Foo, Melanie Fan, Will Holdsworth, Teresa May Koh, Nat Cortazar, Joana Rocha, and Aditi Mishra Feels good to be home. P.s shoutout to @leo1yu for helping make this happen 🙏
12
11
1
41
4,384
In January 2009, a plane full of people lost both engines about 2,800 feet above New York. No runway in reach. 155 people on board. Freezing water below. Every one of them survived. The world called it a miracle. The captain had a very different explanation:
1
10
292
4,508
AI Archive retweeted
In May we published a result that got almost no attention. It was not a drill intercept. It was metallurgical testing on the disseminated, lower grade material at Nisk. The high grade intervals are what people quote back to me. I understand why. They are the numbers that stand out, and they are the reason most people looked at this project in the first place. But a deposit is not only its best metres. Between the high grade zones there is a lot of lower grade material. Whether that material responds to processing is not a detail. It changes what you are actually looking at, because it determines how much of the system is worth anything at all. That testing came back strong. I would rather shareholders understood this than remembered another headline intercept. The best holes tell you something is there. The work on the material around them tells you whether it adds up to a mine. If you only ever read the headline intercepts, you are reading half of it. $PNPN.V | $PNPNF | $IVV1
2
10
28
2,009
AI Archive retweeted
When a Chicago porch collapsed in 2003, officials first blamed the crowd. Then inspectors looked at how it had been built. It was the early hours of June 29, at a party in an apartment building on West Wrightwood Avenue in Lincoln Park. The third-floor deck gave way onto the one below. Thirteen people died and 57 were injured. The first explanation was the one that is always available. Too many people, some of them jumping. The investigation told a different story. The city charged that the porch jutted out further than the code allowed and had been built without a permit. Inspectors found it lacked proper supports, had undersized wood flooring, and was fastened with screws too small for the job. A crowd is a load. A porch exists to carry one. The litigation took ten years and ended in a $16.6 million settlement. The city tightened its codes and started inspecting far more porches. Chicago recorded 6,670 porch violations in 2006. By 2019 that was down to 2,127. I have presented this case to Texas lawyers in a State Bar CLE on catastrophic collapse, because the pattern repeats. The first explanation after a failure usually points at how the structure was used. The real answer is usually in how it was built. If you own a building and something fails, ask for the construction record before you accept the first story.
1
9
1
17
3,122
Someone made a London tube map for salespeople. Pretty accurate, except there should be about 67 more follow-up central stops before you reach closed/won. Love the creativity from @smartlead_ai
1
10
21
44,869
AI Archive retweeted
Some founders go on podcasts. Some write thought leadership posts. @smartlead_ai put theirs on a Times Square billboard and made them dance. I think we've officially reached peak founder-led marketing.
1
12
14
35,635
AI Archive retweeted
Imagine spending all day worrying about your sales pipeline, only to have the train station bring it up on your way home. Even Wynyard Station won't let you clock off. @smartlead_ai
9
11
34,867
AI Archive retweeted
In 2020 I founded a campaign to stop people selling shares that don't exist. Started it as Save Canadian Mining. It runs as Save Canadian Markets now. It exists because of a rule change most investors never heard about. October 2012, IIROC removed the tick test. After that you could short a stock on a downtick, repeatedly, without ever locating the shares. For a large cap with deep liquidity that's noise. For a junior explorer with a thin float it's a weapon. A junior funds itself by issuing shares. If the price can be walked down on demand using paper that was never located, every financing is more dilutive than it should've been. The company issues more shares for the same dollars. Existing holders get diluted for reasons that have nothing to do with the rock. That's a transfer from people who took real risk to people who took none. I'm not asking for sympathy. This sector has plenty of bad operators and the market's entitled to punish them. What I want is price discovery that reflects what a company actually found. Our resource estimate published on 8 September. The first deep hole assays came out on 22 September. Both get priced by a market I still think is structurally tilted. So we do the only thing in our control. We publish, we explain, and we let the drilling speak. $PNPN.V | $PNPNF | $IVV1
4
16
35
2,304
Michael Connerley bought a cafe with no staff and learned every job in it in a month. Then he stepped away from it for 6 months to run someone else's. Here is his story:
1
9
11
5,086
AI Archive retweeted
For 15 years, America used about the same amount of electricity every year. This year it will use more than it ever has. That is the Energy Information Administration's forecast from earlier this month: 4,135 billion kilowatthours in 2026, a record, rising to 4,211 billion in 2027. The agency names two drivers. Data centers, and manufacturing. The turn came quickly. After a decade and a half of nearly flat consumption, demand has grown 2.1% a year over the last five. In January, EIA noted it was the first time since 2007 that power demand had risen four years in a row. Texas felt it first. On July 22, ERCOT set a new all-time peak of about 91 gigawatts, clearing the old record from August 2023 by more than 5 gigawatts. In 2021 I gave a talk called A Ruins to Renaissance Playbook, about reframing America's infrastructure. One of the last slides carried three words. Resilient Electrical Grids. At the time, demand had been flat for over a decade and the grid was treated as a finished system. Roads, ports, water and rail each had a moment when the country chose to build ahead of demand. Electricity has reached that moment again.
4
9
18
1,825
AI Archive retweeted
I think you're missing the point young sirs. We had our intern go and make the real chart of what really matters Tis not how many pints one drinks but how fast one drinks them @destraynor nabs it on this one with 1 pint per 43 minutes, sorry @elonmusk
21
11
7
145
107,474
1 in 5 Australian women aged 45 to 54 is an unpaid carer. Most carers of working age are holding down a job too. Here's what the numbers say about the sandwich generation, and the one thing the research says to watch: At Xplor, we built software for three people: the parent, the educator and the childcare business owner. Ten years later, those three are the same person. Caring up for a dependent parent and down for a dependent child. That's the sandwich generation. The ABS last counted carers in 2022: - 3.0 million Australians were unpaid carers, up from 2.6 million in 2018 - 1.2 million were the main carer for someone, and two in three of them were women - Of those caring for a parent, almost six in ten were aged 45 to 64 - 70% of working-age carers had a job, against 79% of people who weren't caring Most people assume all that caring wears your health down. The research is more specific than that. A 1999 US study followed older people caring for a spouse. The carers who said they felt strained had a 63% higher risk of death over four and a half years. The carers who didn't feel strained had no raised risk at all. A bigger study in 2013 matched 3,503 family carers against 3,503 similar people who weren't caring. Over six years, the carers had an 18% lower death rate. But inside that group, carers under high strain had a death rate about 1.5 times that of carers with no strain. The risk followed the strain. These are observational US studies, mostly of older carers, so they don't prove cause. They do point at the same thing: how much you're carrying, and whether you're carrying it alone. It takes a village to raise a child. It takes one to care for a parent too. If you're in the middle of it: - Be honest about how much you're carrying - Tell someone, even if nothing changes straight away - Hand one job off this week, however small If you know someone in it, ask them what you can take off their plate. Follow me @SammyRArmstrong for more on health, community and building something that matters.
2
9
12
1,566
AI Archive retweeted
Five of the metals in our new resource sit on the critical minerals list of four different governments. Copper. Nickel. Cobalt. Platinum. Palladium. Quebec, Canada, the United States and the European Union all four list them. Silver makes the American list as well. That is not a quirk of geology. It is what happens when the same supply problem turns up in four capitals at once. For us it is practical rather than patriotic. Canada has a refundable tax credit worth 30% of the capital cost of eligible new equipment for extracting and processing copper, nickel and cobalt. Quebec has a refundable credit of up to 45% on exploration and development spending for critical and strategic minerals, a ten year tax holiday for large investment projects, and a 2.5 billion dollar fund for the sector. We believe we may qualify for several of them. Nothing is guaranteed. Eligibility is decided by the programs and the legislation behind them, not by us. Governments are investing directly, via favourable loans, direct grants, tax credits and now accelerated deductions. This is unprecedented. They see the urgent need and are trying to help kick start the sector.
3
10
1
30
2,213
If you have posted about NDIS rorts and never been near the scheme, I have an invitation. Come to our awards in November. Sit in the room. Look at the work. Then go home and post whatever you like. The media angle on the NDIS is fraud, waste and rorting. Some of that is real and it should be fixed. But it has become the only story we tell about the scheme, and it is not the whole story. Not every person in the NDIS is fraudulent. Most of the people in it are doing good work for very little, and you never read about them. So we made an awards night for the other side of it. The people, the providers and the ideas that are actually improving how care works in this country. It is on Thursday 26 November at the Hyatt Regency in Sydney. Black tie, which is the part I like least. My speech last year was about 3 seconds long. Welcome, have fun, sit down. It will be the same this year. We invite everyone. Providers, support workers, participants and families, public servants, ministers, media. If you write about this sector, come and meet it. Come along, see the work, and if you still think it is all a waste of your taxes, post that. At least you will have seen it first. kismet.healthcare/awards
9
12
1,222