Founder Archive

Walt Disney got fired for "lacking imagination." James Dyson created 5,127 failed prototypes. Sara Blakely was told no by every manufacturer. Elon Musk nearly went broke funding Tesla and SpaceX. Jack Ma failed his college entrance exam twice. Here's the truth nobody admits: Every founder you admire failed their way to success. The only difference? They understood failure is tuition, not a verdict.

Yvon Chouinard spent fifty years refusing to run Patagonia like a business. He capped growth, told customers to buy less, and gave it away at 83. In five years not one mother has failed to return from maternity leave. Here's what he did that most leaders get backwards:

SpaceX is one of the most impressive companies on the planet. I genuinely believe that. But Facebook was also an incredible company when it went public at $104 billion. The stock still dropped 54% in four months. Anyone who bought on listing day spent over a year underwater. SpaceX just went public at 17 to 20 times that valuation. Great company and great entry point are two completely different conversations. The best risk-reward right now is in assets the market hasn't discovered yet, where the upside is still ahead of you rather than already priced in. A high-grade critical minerals deposit with 20x the global average copper grade, 95%+ recoveries, and a Mineral Resource Estimate weeks away, trading at a fraction of NAV. That's where my own money is. $PNPN.V | $PNPNF | $IVV1

In August 2008, one corrupted database stopped Netflix from shipping DVDs for three days. Their answer took seven years, replaced nearly every system they owned, and ended with software that kills their own servers at random. Here's the rebuild, and the drill any team can copy:

Tahlya Sleep was 25 when a 13 year old in her care asked if he could come with her. She was moving interstate. He was staying in the child protection system. She still checks in on him today. Here is her story:

Your insurer may have photographed your roof before the last storm hit. Carriers across Texas are now using satellite imagery, drone overflights, and AI to document the condition of properties in real time. High-resolution photos are timestamped, geolocated, and stored. When you file a claim after a hailstorm, your insurer pulls up images from six months earlier. A curling shingle. Moss. A stain. Anything that suggests the damage was pre-existing. Contractors submitting claims with what insurers call "ambiguous damage evidence" are seeing 30 to 50% higher denial rates. The technology works. The way it's being deployed is the problem. Insurers are building a pre-loss record that justifies post-loss denials. The assessment happens before you even know there's damage to assess. Property owners can take one step right now: document your property's condition independently. Hire your own inspector. Get a dated report. If your insurer shows up with drone footage, you need your own baseline - not theirs.

A boy in a fog covered Kent churchyard is grabbed by an escaped convict, and his life bends around it. A young officer rides into a desert that has swallowed armies and comes out a legend. Great Expectations (1946) and Lawrence of Arabia (1962).

At mile 22 of the London Marathon, the crowd started shouting at Gordon Ramsay about his weight. He was 34 years old and weighed 108 kilos. Twelve years later he finished the Ironman World Championship in Hawaii. Here are the 7 things he changed:

Google Classroom now lets a teacher assign an AI agent to a single student. Not the class. One student. I sit on the board of my kids' school, and that one feature changed what we're planning for. Here's what school looks like in 5 years:

Something worth watching that hasn't gotten much attention yet: Nisk Far West. After our 2025 land acquisitions, a copper soil anomaly was identified sitting right next to interpreted ultramafic rocks — the same rock types that host both the Nisk and Lion deposits. We drilled a single diamond hole into the area. It confirmed ultramafic rocks. Now we're deploying a full Ambient Noise Tomography survey - 200 sensors across a 14km by 3km grid - to build a 3D model of where Cu-PGE and Ni-Cu zones may sit within this new target. Assays from the winter reverse circulation program are still pending. High-priority targets get follow-up diamond drilling. This is how a property evolves from known deposits into a district. Acquire the right ground, apply the right science, let the geology direct the next move. The Nisk project keeps getting bigger. Worth watching as results come in. $PNPN.V | $PNPNF

Zillow spent a decade building a model that could price almost any home in America. Then it let that model buy houses. Two quarters later it wrote off about 550 million dollars and shut the business down. Here's what broke, and the question to ask before a model spends:

John Timpson built one of Britain's best known high street chains. Nobody there is told how to do their job. No scripts, no sign off, almost no rules. Most owners would call that chaos. He calls it the whole point. The 2 rules the whole chain runs on:

It takes most teams 2+ weeks to send the first cold email from a new domain. Ours go out from day 1. Warmup is the tax nobody prices in as a new domain has zero reputation, so Gmail treats it like a stranger. This is usually a pain because you sit there for 2 weeks sending fake warmup emails back and forth just to look trustworthy before you can launch the campaign. It gets worse the more inboxes you need. You either wait 2 weeks, or rush fresh domains and eat the spam folder. We stopped warming from scratch a long time ago. This is our setup: - buy every inbox pre-warmed. It shows up with months of real sending history and DNS already handled - cap it at 15 to 20 sends per inbox per day, forever. Low volume per sender, spread wide - leave warmup running in the background even while campaigns send - watch health per domain and pull anything that slips before it drags the rest down A campaign approved on Monday is sending Monday. @smartlead_ai made this native recently btw. Pre-warmed mailboxes through SmartSenders, delivered in about 30 minutes, 3 mailboxes per domain, sending the same day. Send from day 1 with Smartlead’s pre-warmed mailboxes

The warmest lead in your pipeline never replied to you. Go look at your own sequences and you'll find them - someone who opened your email four times, read every follow-up, and never said a word. Most teams write that person off as dead and go dig up a fresh stranger to email instead, which is probably the most expensive habit in outbound. A lead who keeps coming back to your email is already circling. They know your name and your offer, and the only thing they're missing is a reason to talk to an actual human. That's why reps who add a call step to their email sequence see up to 2.5x more positive replies: the email did the slow work of building familiarity, and the call just walks in and collects it. So why does almost nobody pick up the phone? It comes down to context. By the time a rep is ready to dial, the story of that lead is scattered across the CRM, the inbox, and the sequence tool, and rebuilding it before every call is enough friction that the call quietly never happens. That's the part @smartlead_ai's new SmartDialer actually fixes. It lives inside the email sequence itself and drops the whole thread on screen before you dial, every send, every open, every reply, so the call picks up exactly where the email left off instead of starting from zero. It transcribes and logs the call after you hang up too, so the follow-up doesn't ride on whatever the rep happens to remember. The email creates the signal and the call is what converts it, and the teams that win the next few years of outbound will be the ones who wire those two together on purpose instead of treating the phone as an afterthought. How many "dead" leads in your pipeline are really just waiting for a phone to ring?

Most of my cold email skill is just knowing where to click when something breaks. When a campaign stops sending I don't panic anymore, I just work through the same routine: check the schedule, then every mailbox connection one by one, then the warmup scores, then the sending limits. Nine times out of ten it's one of those five things, and I only know that because at some point I've been burned by all five of them. It took me years to build that instinct, and I built it the worst possible way, by breaking something and then working backwards to figure out what I'd broken. Emails quietly slipping into spam cost me a week of my life once, and a campaign that looked completely live but wasn't actually sending cost me another. Nobody really teaches you this part, because everyone's busy talking about copy and targeting when the thing that actually decides whether you survive cold email is whether you can read DNS, deliverability, warmup and sending limits when they go sideways. You only learn where to look after it's already cost you a few deals. That invisible routine is the real learning curve, and it's the reason most beginners quit before they ever see a reply. They hit a wall they have no way of reading and decide the whole channel is broken, when really they just didn't know which of the five things to check. I've thought more than once about how much faster I'd have gotten good if I could have turned to someone who could actually see my account and just asked them what was wrong. @smartlead_ai's SmartAssistant is the update I'd have wanted three years ago. It's a chat box in the corner of every page. You type "why aren't my emails sending?" and it reads your actual account, your campaigns, your mailboxes, your history, then hands back the specific cause and either fixes it or walks you through it. Same for "create a campaign" or "show me my analytics", it acts on your account instead of linking you to a help article. The routine that took me years to memorize is now something a brand new user gets to skip entirely. It's live in the bottom-right corner of Smartlead. Next time you'd reach for a support ticket, try just asking it first.

Charlize Theron is 50. In Apex she spends the film hanging off cliffs in the Blue Mountains without a harness. At 29 a stunt left her one centimetre from being paralysed. Eight years of nerve damage followed, then surgery on her neck. Here are the 6 things she changed:

In 2022, attackers stole customer data from LastPass by breaking into one engineer's home computer. The way in was a media app on that machine that had not been updated in over two years. Here's how a home movie server reached the keys to millions of password vaults:

Sterling Hayden had a face like a cliff and a gift for playing men whose perfect plans fall apart. He played that man twice, for two different masters, in the two films that built the modern heist movie. The Asphalt Jungle (1950) and The Killing (1956).

10 million more people will call Texas home by 2060. Over 90% of them will settle inside the Texas Triangle. No other state in America has this kind of runway. The opportunity is massive - if we build to match it.

Power Metallic Update from Sydney MRE delayed to end of August due to Engineering workload Company is on the road prepping market for a positive MRE Followed by Summer drilling assays later in September Looking for a strong close to 2026 $PNPN.V | $PNPNF | $IVV1