WATCHMAN'S TORCH Issue #32

Reflecting on my childhood in Karachi.

By Gabriel Custodiet

Downloading a new 10 GB game release was, at best, an overnight affair. On a good week, PTCL’s DSL (Pakistan Telecommunication Company Ltd) would hold at 1 Mbps and the download finished overnight. On a normal week, it was a two-to-three-day project — pausing during the day so the household could use the line, resuming after midnight like a smuggler working by moonlight.

Except I had a second enemy no congestion graph could show: the electricity. This was the height of load shedding, and KESC (Karachi Electric Supply Company Limited) cut the power for hours a day, often without warning. Many a night I woke to a dead screen and a download frozen at 40%, murdered by a 3 a.m. blackout. The DSL exchange usually had generator backup. The internet was still out there, humming patiently in the dark and every serious household owned a UPS. Ours powered the router and nothing else.

Part of the problem was geography. At the time, Steam had no CDN (Content Delivery Network) in the entire region. Every byte had to crawl across congested international links from far away, and every release day turned into a tug of war between a million frustrated teenagers and a handful of overloaded transoceanic cables.

“Why not just buy the DVD?” Because Rainbow Centre in Saddar sold cracked discs for fifty rupees, and cracked discs don’t get you online. Online meant the latest version, and every patch was another night of downloading. Legitimate retail barely bothered with us, and even when it did, Steam wanted a credit card the local banks wouldn’t issue and dollars at whatever ruinous rate the rupee had fallen to that month. The gray market of wallet codes and fifty-rupee discs wasn’t a lifestyle choice. It was the market.

Meanwhile, my mother would appear at 8 p.m. sharp, demanding to know why the drama serial (a type of local telenovela often in Urdu) was “acting up.” The serial ran on cable television. It shared no wire, no box, no single electron with my download. No amount of explaining this ever saved me.

I desperately wanted more. Faster. So I did two things.

First, I bought my first 1 TB hard drive — a small fortune at the computer market, and more storage than I believed I could ever fill.

Second, I resurrected an old laptop and turned it into a Squid proxy (yes, that is really the software’s name). I rerouted the entire household’s traffic through it, man-in-the-middle style: every request passed through my little machine first, and anything already in the cache was served instantly instead of being dragged back across the ocean. The whole house got faster, and nobody knew why but me.

The laptop choice carried a second, accidental genius: its battery was a built-in UPS. When KESC cut the lights, my proxy kept running.

My favorite ritual was watching the hit rate — the counter showing how many requests my cache had answered without touching the outside world. Every hit was a small victory over the last mile, the rationing, the blackouts. I was unreasonably proud of that number.

If you’ve been reading a while, this playbook should sound familiar. It’s the same trick I deployed years later as a student working for an ISP on a small island in the middle of the Indian Ocean TO THE READER. Watchman’s Torch Issue 006, where a 20-terabyte YouTube cache stood between us and a sluggish satellite backup link. Same playbook, different scale. Pakistan would later become famous for accidentally knocking YouTube offline for the entire planet but that is another issue entirely.

Why am I telling you this? Because we sell a course called Internet Anywhere — and it is not “buy Starlink and pray.” It’s decades of strategy earned on the most adversarial connections you can imagine: islands, mountains, basements, blackout grids. All in service of one goal: internet that just works, anywhere.

On Monday the 17th at midnight, the price rises from USD 399 to USD 499. This is your last chance at the current rate. Can’t attend live? Every session is recorded, and the recording is yours either way.

The 2026 Coldcard incident was a black smear on the final days of July.

By Gabriel Custodiet

We covered the massive theft due to poor generation of seed phrases by the Coldcard random generator on [Episode 229 of The Watchman Privacy Podcast – Coldcard: The Devastation Begins]

This incident led to a lot of us scrambling. We were helping lots of people: friends, clients, and our audience as they tried to move their funds securely before they could become part of the $100+ million in stolen funds (and growing). To be fair, we haven’t recommended Coldcard wallets for many reasons for years, but it was a popular choice for many and so we joined the front lines in helping people move their funds out of these wallets.

The truth is—and I’ve always expressed this opinion when the topic arises—I don’t really believe in crypto hardware wallets. And I’ll explain why in a moment. But I’ll also say for starters that I know a lot of hardworking, entrepreneurial, and privacy/sovereignty minded people who work at these companies. Aside from Coldcard, they’re making an honest living. And I do think they produce a product that can have strategic value.

My job is not to defend any company or organization but to state my opinion with your best interests in mind, dear reader. If I had a company sponsoring me, of course, I would have a different opinion and would mentally justify it to myself.

I recently helped someone to transfer funds due to Coldcard vulnerability. We were about to set up an elaborate new hardware wallet that worked best with its own software: two new pieces of technology to get familiar with, and plenty of room for mistakes or procrastination and making sure the “right version was installed.” That’s when he asked me: “What is the point of a hardware wallet?”