I Need Wisdom: A Letter, Fifteen Years Later

Where it starts: a letter, fifteen years before the first line of code

On 12 June 2011, PJ's father, Kobus, wrote him a letter. Kobus had been diagnosed with stage 4 melanoma; PJ, knowing there was a real chance his father wouldn't survive it, had asked him for his wisdom while there was still time to ask. Kobus wrote back by hand, his handwriting visibly unsteady from the illness. He died eight weeks later, at 50.

The letter covered the ordinary, hard-won things a father tries to pass on before he runs out of time: humility, sticking with a decision once it's made, providing for family, giving generously. One line from it has become the project's public touchstone: "Remember above all: it's God's money and God's world — we're only here a little while to enjoy it." PJ has said plainly that he wishes he'd asked his father sooner, and asked for more.

That wish, sitting unresolved for fifteen years, is the actual origin of I Need Wisdom: a project to sit down with people 80 and older while there is still time, ask what they've learned, and give it away for free so nobody else has to wish they'd asked sooner.

Phase 1 — Laying the foundation

The planning session that produced the project's master plan locked in three decisions: remote-first interviews over video call rather than starting with in-person visits; no payment and no product to sell, so the question of exploiting an elderly guest never has to come up; and a named risk addressed early — that the easiest sourcing channels skew toward accomplished, already-online people in their 60s and 70s, which could quietly turn the project into "career advice from retired executives" instead of its intended, more distinctive shape. The countermeasure was to keep 80+ as the published bar and lean on referral chains and community groups to find more varied voices.

Phase 2 — Claiming the name

This phase produced the domain, a branded inbox forwarding into a personal account, a cover-art spec, a trailer script, and a plan to claim identical handles everywhere. The preferred handle was already taken, so the project settled on a close variant across every platform. One honest limitation surfaced immediately and stayed true for the rest of the project: an AI assistant cannot create social accounts, enter payment details, or solve CAPTCHAs. Every account had to be created by hand, with all the copy prepared in advance so setup was closer to copy-paste than to writing from scratch.

Phase 3 — The origin story becomes the mission

This is where the letter got transcribed and translated, and where the mission statement — "I Need Wisdom exists so that no one has to wish they'd asked sooner" — got written directly from its lines. The real decisions in this phase were about restraint: which single line from the letter to make public, whether to show a photograph of the actual letter (yes, with the section where his father writes more personally about love gently blurred, keeping the practical lessons visible and the most private part private), and confirming his father's full name for a proper dedication. That memorial section didn't actually reach the live site until much later — it sat prepared for weeks before it was deployed.

Phase 4 — Launch mechanics and the first outreach drafts

A consent form went live, built to route signed releases into the site's form handler alongside the nominate and waitlist forms. Getting from "form exists" to "form actually captures submissions" took more than one pass — form detection had to be explicitly turned on and only took effect on the next deploy, and partway through, the hosting account ran out of build credits, blocking new deploys entirely until it was topped up. Once restored, all three forms showed as live and collecting, checked directly in-browser rather than taken on faith.

This phase also produced the first round of outreach drafts to a short list of well-known potential guests. A useful, slightly deflating piece of honesty came out of this work: well-known people mostly don't have public personal email addresses — reaching them realistically means going through a manager, foundation, or publicist, and for at least one, a personal social-media message rather than email.

Phase 5 — Expanding the guest list, finding real doors

The bottleneck named directly was "finding guests," so this phase went wide on research: around 55 new prospects were added, organized by how reachable they actually were rather than by fame alone, with real, verified contact paths located for several. Two vetting habits got established here that recur through the rest of the project: a hard exclusion bar (anyone with an active lawsuit or unresolved scandal is left off the list entirely, regardless of fit), and a standing mortality check — because many ideal guests are in their 90s and 100s, the guest-research files carry a permanent rule to re-verify anyone 80+ in the same week an email actually goes out.

Phase 6 — One especially personal piece of outreach

PJ found a 1991 photograph of his father standing with a golf legend both of them admired at a golf day. He also remembered, from a book he read after his father's death, that golfer's word for the tradition of his family and farm staff sitting together in a circle: an "indaba." Those two threads became the spine of the project's most personal invitation so far — refined over several rounds toward a warmer, plainer tone, with the interview length deliberately left open-ended rather than boxed into a fixed slot. The old photograph needed real restoration work: cropping out its frame and mat, correcting decades of color fade, and producing a captioned version identifying who's who and when it was taken. As of the most recent work on this thread, the message was refined and ready — but not confirmed sent.

Phase 7 — Building the Wisdom Library

This turned into the single biggest engineering effort of the project so far: a collection of 1,114 King James Version verses across Proverbs, Ecclesiastes, and Song of Solomon, pulled programmatically and cross-checked against a second independent copy; a collapsible list of principles paired with modern-day application and linked verses, launched with 20 and expanded to 40; and four more dossiers summarizing the project's other intellectual influences.

Two real problems showed up along the way. First, a request to switch the Bible text to a more readable modern translation ran into a genuine legal wall — modern translations are copyrighted and prohibit republishing whole books, which is exactly what these pages do. The resolution was to keep the public-domain text on-site and add "easier reading" links out to a Bible-reference site for the modern version of every passage, explained openly as a licensing choice. Second, a technical problem showed up here and kept recurring for the rest of the project: the working copy of the site would sometimes lag behind real edits, serving stale or corrupted copies of files that looked fine on the actual source — at one point producing a deploy where the homepage appeared completely blank. The fix, each time, was some combination of waiting for sync, writing to brand-new filenames instead of editing in place, and verifying every file before every deploy — a discipline that became standard practice from here on.

Phase 8 — Three compilation videos, researched from scratch

Three video research kits got produced — short, fair-use compilations of well-known people saying some version of the project's core wisdom, as a way to reach an audience before the interview pipeline is producing episodes. What's notable is less the concept than the rigor built around it: every quote is tiered by confidence rather than presented as uniformly solid, each kit carries its own "misattribution minefield" of widely-shared quotes that turned out to be fake once checked, and a standing exclusion rule kept several otherwise-perfect candidates out over active controversy. One deliberate ethical line, separate from any legal one: a well-known line about loneliness, sourced from an interview the broadcaster obtained through documented deception and has since pledged never to air again, was left out entirely even though the quote itself is real.

Phase 9 — From research to a rough cut, and the daily-blog engine gets built

A first rough cut was assembled purely to check pacing, then a companion blog was built connecting daily news to the wisdom library. Direct, useful pushback shaped both: the first post read as visibly AI-written, and a phrase the project had been repeating was the wrong one to keep using. Both got fixed — the post was rewritten in a plainer, more human register, and the site adopted a new framing phrase specifically requested: "let's see what wisdom says." A second rough cut fixed a cut-off-mid-sentence problem by re-trimming every clip using audio silence-detection rather than trusting auto-caption timestamps, which turned out to lag real speech by up to a second. The session closed by writing a full "master prompt" codifying the whole daily process, and a recurring scheduled task was set up to run it automatically every morning.

Phase 10 — The origin story goes live, and the daily habit runs on its own

The memorial section drafted back in Phase 3 finally reached the live site — the story, the softly-blurred letter photo, and the dedication — checked by reloading the live page rather than trusting that the deploy had reported success. Social profiles got tidied up around the same time. And the daily post went from a one-off to an actual habit: the scheduled task ran on its own for the first time, unattended, picking a real current event and reading it against the project's own sources — and, mid-publish, caught a live version of the Phase 7 sync bug and worked around it using the same verify-before-deploy discipline that had by then become standard.

Where things stand right now

Live: the website, the full Wisdom Library, a working consent and nominate pipeline, a YouTube channel, an Instagram profile, an X profile, and a daily post now running on an unattended morning schedule. Built but not yet published: a second rough cut of the first compilation video, still missing some clips, and complete research for the other two with no footage assembled yet. Ready but not yet confirmed sent: several personalized outreach drafts. Still blocked on PJ specifically: recording the actual podcast trailer, finishing the remaining social profile setups, and creating the podcast-hosting account itself — all steps that require his own login or his own voice, which is exactly as it should be.