We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
The standout recommendation
Title:To FDE, or not to FDE?
Type: Article
Creator: @thejessezhang
Recommended by: Aaron Levie (@levie)
Levie calls it a “good post” because AI introduces a non-deterministic, rapidly changing system into workflows that often have never been automated. He highlights the resulting need for customer-process change, customization, constant evaluations, model updates, and iteration.
The essay’s practical framework is to deploy engineers directly with customers when a category is genuinely new: observe failures, discover the workflow, and turn that pain into product. Once the user journeys are understood, companies should remove the FDE dependency; otherwise bespoke work keeps margins capped and growth tied to hiring. The useful distinction is between FDE-led discovery and implementation against a known specification. The article’s test is blunt: use FDEs to determine what the product needs to become; if they keep excreting more pain, the company has a services business rather than a product motion.
Tim Ferriss’s 80/20 diagnostic
Title: Pareto’s work—the 80/20 lens
Type: Framework / work
Creator: Pareto; Ferriss does not identify a specific title or edition
Link: No link supplied in the post
Recommended by: Tim Ferriss
Ferriss recounts discovering Pareto’s work while working 15-hour days and facing burnout. He then audited his business and personal life with two questions: which 20% of sources create 80% of his problems, and which 20% generate 80% of his desired outcomes?
The recommendation is valuable as a usable reading lens rather than a generic productivity slogan: eliminate the sources producing disproportionate harm and multiply the strengths producing disproportionate results.
Paul Graham’s creator recommendation: Rob Miles
Title: Rob Miles’s X posts—illustrated by “Never pay someone to remove a problem that they themselves created”
Type: X posts / creator to follow
Creator: Rob Miles
Link:Example post
Recommended by: Paul Graham
Graham says Rob Miles is worth following because he delivers insights that are both “very general” and “novel,” then points readers to the example principle above. Graham qualifies the endorsement: novelty is relative, but the principle is not widely appreciated. That gives readers a clear filter for the account’s value—look for operating principles that generalize beyond one situation without collapsing into familiar conventional wisdom.
Kristen Lowe (executive ghostwriter/narrative strategist, incoming Director of Founder & Editorial Communications at Scribe) guest-authored this Lenny's Newsletter guide to founder-led communication; it explicitly recommends the following non-commercial resources as exemplars by archetype:
- Podcast: Sara Blakely on Masters of Scale (episode) — 'a detailed origin story' template for Problem founders
- Podcast + article: Dave Gilboa (Warby Parker) on the Good Life Project and his Daily Californian interview — telling the same story across mediums
- Podcast + blog + tweet: Will Ahmed (Whoop) — How I Built This episode ('well-rehearsed but not canned'), Locker Blog post, and tweet for short written storytelling
- Podcast: Tope Awotona (Calendly) on How I Built This — contrast with money-driven startups shows why personal narrative works
- LinkedIn posts: Sara Blakely's posts and birthday post — 'the template' for problem-storytelling
- X: Aaron Levie (Box) — posts and original insights as model Insight founder; his IT-leader learning series is 'the best example' of helpful vantage point
- Article + site: Patrick Collison (Stripe) — The Atlantic byline 'We Need a New Science of Progress' and personal website — thought leadership beyond company talk
- Video + posts + essay: Jason Fried (37signals) — TED Talk 'Why work doesn't happen at work' ('brilliant example of explaining without patronizing'), X posts, and Signal v. Noise essay
- Blog: Intercom Blog, e.g., Des Traynor's 'Product Strategy Means Saying No' — strong company-hosted thought leadership
- Article: Rahul Vohra (Superhuman) — First Round Review piece on the PMF engine — 'standout example of letting people into your reasoning'
- Audio: Steve Jobs' internal 'Think Different' talk (skip to 6:30) — clarity on what you stand for
- Letter + video: Melanie Perkins (Canva) — 'note to the Canva Community' (big announcement without making it about you) and SXSW interview
- Essays: Elon Musk (Tesla) — Master Plan Part I and Part II — making vision concrete without sounding naive
- Essay: Yvon Chouinard (Patagonia) — 'Earth is now our only shareholder' — top-tier 'x doesn't have to be true' messaging
@levie recommended @thejessezhang's article “To FDE, or not to FDE?” as “Good post how to think about FDEs” (x.com article: https://x.com/i/article/2086604287136980992). He argues FDEs are “real and not going away for AI any time soon” because AI introduces a non-deterministic, rapidly changing system into workflows that have largely never been automated, unlike deterministic software; agent deployments require changing customer processes, heavy customization, constant evals, model updates, and iteration from feedback . The recommended essay argues for forward-deployed engineers early to discover workflows in new categories, then removing them as the product absorbs the findings; it repeats Palantir CTO Shyam Sankar's “FDEs eat pain and excrete product” and warns that FDEs producing “more pain” mean a services business, not a product .
Paul Graham recommends following Rob Miles (https://x.com/robertskmiles/status/2087235360690741690), calling his insights "very general, but also novel" – "the rarest thing of all" . Example of Rob Miles' thinking: "Never pay someone to remove a problem that they themselves created," shared in a tweet linking to Sam Altman's post . Graham adds that novelty is relative – similar ideas exist in history, but this principle is "not a widely appreciated principle" .
Recommendation: Aaron Levie endorsed the X article "To FDE, or not to FDE?" by @thejessezhang — "Good post on how to think about FDEs" — with a link to it . URL: x.com/i/article/2086604287136980992 .
Why Levie recommends it: He argues FDEs are "real and not going away for AI any time soon" because AI is about adding a non-deterministic, rapidly changing system to workflows that have largely never been automated before, unlike deterministic traditional software; heavy customer-process customization, constant evals, and ongoing model updates mean this work remains — and may even get more complicated — as enterprises throw increasingly complex processes at agents . He specifically highlighted the article's observation that for an AI agent for accounting in 2026, no established workflow exists: "Nobody knows what the user journey looks like — not you, and importantly, not your customer either. They can't tell you what they want, because the thing they'd want doesn't have a shape yet" .
Core framework from the article: Use FDEs early to discover workflows, then productize and pull them out — "FDEs eat pain and excrete product" ; the trap is not starting but not stopping ; and FDEs are not the same as implementation . The article concludes: "Go forward-deployed early. Get the signal. Put your engineers in front of customers forever" then ask whether the bespokeness is in the customer's environment or your product gaps .
Tim Ferriss recommended applying Pareto's work (the 80/20 principle) after discovering it during a period of burnout from 15-hour work days . He used two diagnostic questions — 'Which 20% of sources are causing 80% of my problems and unhappiness?' and 'Which 20% of sources are resulting in 80% of my desired outcomes and happiness?' — to analyze his business and personal life . He framed the goal as finding inefficiencies to eliminate them and strengths to multiply them .