We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
The clearest operating case
The High Flyers Podcast #276: John Willett, Co-Founder & COO of Rogo
- Type / creator: Podcast episode from The High Flyers Podcast, featuring Rogo co-founder and COO John Willett.
- Recommended by: Keith Rabois, who posted, “Recommend listening!”
- Key takeaway: The episode follows Rogo through two years with almost no external validation, investor skepticism that the market was too small, early banking customers, and growth from roughly 40 to 200 employees across the US, Europe, and APAC.
- Why it matters: It gives founders a concrete operating sequence—weak early validation, customer traction, then scaling—instead of presenting an AI company only through its eventual outcome.
Two unusually concrete resources
DLSS 5 Autopilot
- Type / creator: GitHub project and Windows utility by Kizzuwatnaa, identified in the repository as its author, reviewer, approver, and project owner.
- Recommended by: Tobi Lutke, who said he had tried it with Cyberpunk, found it “Absolutely incredible,” and thought it was packaged perfectly.
- Key takeaway: The README describes a one-executable tool that adds unofficial DLSS 5 neural rendering to games that did not ship it: it scans a library, selects a route for the game and graphics card, fetches components from their publishers, writes the configuration, and can remove its changes.
- Why it matters: This is a runnable technical artifact rather than a general capability claim. The repository also supplies an important limit: RTX 20/30 community builds are described as heavy, costing about half the frame rate at full model resolution.
Any Human Ever
- Type / creator: Interactive educational web experience. The visible project page does not name a creator; it says only that it was made with the FLUX.1-schnell model on a laptop.
- Recommended by: Gena I. Gorlin called it the most fantastic pedagogical tool she had encountered and urged others to try it; Tobi Lutke amplified the recommendation as “Profound.”
- Key takeaway: The experience draws a year, place, and life step by step, each taken at random from real data, and frames the exercise against the more than 100 billion people who have ever lived.
- Why it matters: It turns demographic scale into a personal exercise rather than an abstract statistic. Gorlin’s plan to use it with her six-year-old makes the educational use case unusually concrete.
A political-history read
“Now the Third Man – Rudi Dutschke – The Man Who Gave the Long March Its Name”
- Type / creator: Long-form X post by @Kristof_Poland.
- Recommended by: Elon Musk, who described it as “Illuminating.”
- Key takeaway: The post’s argument is that Gramsci supplied the theory while Rudi Dutschke turned it into the actionable slogan of a “long march through the institutions,” with universities, media, schools, churches, and legal institutions as the terrain.
- Why it matters: It is a compact starting point for understanding the phrase as a theory-to-strategy narrative. Because it is an argument-driven X post, it is best used as an interpretive entry point rather than a standalone history.
Direct answer:Any Human Ever is an interactive, data-driven experience about selecting one life from the more than 100 billion people who have ever lived. It guides the user step by step through a year, a place, and a life, with each element taken at random from real data.
- Its subject matter is demographic and geographic: when people were born, how population growth makes recent birth years more likely, where people were located around the world, and what a life looked like “then and there.”
- The supplied excerpt does not name the creator. It only says the project was made with the FLUX.1-schnell model “on my laptop.”
- The bundle contains no endorsement text or attribution to Gena Gorlin or Tobi Lutke, so it does not establish the specific educational experience or subject matter they endorsed beyond the project description above.
Direct answer. DLSS5-Autopilot is a single-executable Windows tool that configures unofficial DLSS 5 neural-rendering integrations for games, including games that never shipped DLSS. It scans the library, selects a route suited to the game and GPU, downloads components from their original publishers at run time, writes the configuration, and can remove its changes.
Creator/ownership: The repository identifies Kizzuwatnaa as the author, reviewer, approver, and project owner. The Autopilot itself is a downloader/configurator rather than the creator of every underlying integration; its credits attribute routes to projects such as DLSS5-Feeder, dlss5-bridge, neural-upstream, and OptiScaler.
Concrete setup that makes the capability tangible: It targets Windows 10/11 and NVIDIA RTX 20-or-newer cards. A user runs
dlss5-autopilot.exe, selects a detected game or chooses a folder, reviews the executable/bitness/graphics API/DLSS detection and proposed route, then presses INSTALL; the tool logs the changes and its uninstall removes exactly what it wrote and restores replacements.Most relevant hands-on capability: For a game with no DLSS, the feeder route builds a DLAA contract from ReShade's depth buffer and shader motion vectors, covering D3D10/11/12, Vulkan, OpenGL, and 32-bit games. That is the concrete result a reader could understand an endorsement to validate: whether the tool can make DLSS 5 work in a particular non-DLSS game without a manual multi-component setup. The source also records an important performance qualification: RTX 20/30 use community FP16 builds described as heavy—about half the FPS at 100% model resolution—while RTX 50 uses NVIDIA's build.
Evidence gap: The supplied repository text does not name Tobi Lutke or document his game, GPU, route, settings, or observed result, so it cannot independently substantiate a hands-on endorsement.
Any Human Ever — educational web tool: Tobi Lutke called Gena I. Gorlin’s recommendation “Profound” and linked to her post. Gorlin describes Any Human Ever as “the most fantastic pedagogical tool” she has encountered, recommends that others try it, and says she is excited for her six-year-old to use it.
Keith Rabois explicitly recommended listening to The High Flyers Podcast, episode #276, featuring John Willett, Co-Founder & COO of Rogo. The episode covers Rogo’s early fundraising and market-validation challenges, winning initial banking customers, scaling from roughly 40 to 200 employees, and operating across the US, Europe, and APAC. [http://curiositycentre.com/the-high-flyers-podcast/ep-276-john-willett-chief-operating-officer-rogo]
Linked X post — “Now the Third Man – Rudi Dutschke – The Man Who Gave the Long March Its Name”: Elon Musk explicitly endorsed the post as “Illuminating.” The post explains how Gramsci’s theory and Rudi Dutschke’s “long march through the institutions” slogan became an actionable strategy, including Dutschke’s synthesis of Gramsci and Mao.
YouTube videos — self-directed learning (Auren, amplified by Garry Tan). Garry Tan shared Auren’s post arguing that a crucial skill is teaching yourself something complicated through a YouTube video or equivalent. Auren illustrates the value with a 16-year-old who independently taught herself guitar on YouTube, attributing the result to intrinsic motivation and extending the lesson to programming, hobbies, and other complex skills. Original post: https://x.com/auren/status/2097035375344799789
The post author enthusiastically endorsed the DLSS5-Autopilot GitHub project, saying they tried it with Cyberpunk, found it “Absolutely incredible,” and that its creators “packaged it up perfectly.”
