#123: High-IQ Marketing Moves, How to Become a Robotics Engineer in 6 Months, Figuring Out Your Next Career Move

Hi All!

We are pleased to welcome you to this week's edition of Undiscovered, a newsletter with exclusive resources and insights expanding from the material found on our main site - becketu.com.

This week we will take a look at how one image can become an editable 3D world, how agents should accumulate skills, a stack of high-IQ marketing moves, and more.

Let's dive in:

Generating & Animating 3D Worlds from Images

Three separate posts this week point to an exciting future in the realm of world building.

An incredibly talented team out of China just released Lumera, a turnkey pipeline that reconstructs an engine-native 3D scene directly from one image. It gives you object-level meshes, editable layouts, engine-native lights, HDR environment lighting, and a scene ready for Unreal Engine 5.

Second, NVIDIA open-sourced MotionBricks, one neural model that generates over 350,000 character motions in real time at 2ms latency, dropping straight into UE5. People in the replies mentioned volume was never the bottleneck in animation, and the hard part is the specific timing a director can picture but not name. Regardless, the opportunity is apparent in game development, especially paired with Lumera.

Third, Token Gremlin open-sourced ABYSSAL, a fully procedural ocean and extreme-weather experience running in the browser, with waves, hurricanes, and tsunamis generated in real time using Three.js and GLSL. The craziest part is how the Gremlin did this all on a GTX 1650, a relatively weak GPU.

Put them together and the direction becomes clear. We are moving from generating images and clips to creating editable worlds and everything that moves inside them

Building the AI Skills Infrastructure

​DAIR.AI surfaced a paper from Google on how AI agents should accumulate skills.

The work separates three things that skill-evolution systems usually collapse into one. Raw execution traces, a persistent wiki of accumulated knowledge, and the executable skills themselves. Experience gets consolidated into the wiki, and every later skill update builds on that wiki instead of on a scattered optimization history.

What does it mean?

Smaller models with evolved skills beat substantially larger models without them.

​Khe Hy named the human side of the same coin. Skills are becoming a problem not because they don't work, but because they are so effective. Teams now run 5 to 20 skills each, but skills are stuck in each team member's individual workflow.

There is no version control, ownership, discovery, or testing. An argument is being made that the average knowledge worker will never use git (or even know how to use it).

The models are good enough that accumulated skill is now the edge, but the tooling to manage, share, and retire that skill does not exist yet. This presents a unique opportunity for anyone looking to build infrastructure for AI skills.

High-IQ Marketing Moves

​Cody Schneider spent the week doing a brilliant X series I could not stop reading. He takes a marketing channel and lays out the low-IQ version, the midwit version, and the high-IQ version. The joke is that the low and the high are usually the same simple thing, and the midwit version is all the expensive busywork in the middle.

Here are the high-IQ moves he shared:

​SEO - Ship a lot of pages that each contain something not available elsewhere your data, your opinion, a tool that works. Match the intent behind the query precisely. Make the site crawlable and fast enough. Get links by making things worth linking to. Ignore rankings, watch revenue.

​AI search - Split it into the two things it actually is. (1) Retrieval: rank in the underlying search index, because that's where the citations get pulled from. (2) Consensus: get named on the third-party pages that get retrieved — Reddit, YouTube, G2, "best X for Y" roundups, other people's blogs. Then publish primary data nobody else has so you're the only citable source for some claim. Measure it with referral traffic and "how did you hear about us."

​Link building - Build things that get cited without asking, proprietary data, a free tool, a strong public opinion. Then do the two unglamorous, highest-ROI plays: get placed on the listicles that already rank for your category, and build integration/partner pages with companies you actually integrate with. Measure referral traffic and rankings, never "links built."

​Facebook ads - one Advantage+ campaign, broad, CBO, optimize for the deepest event that gets ~50/week, and spend all your time making more creative

​Google ads - A handful of campaigns (one or two Search, PMax if the inventory/feed justifies it), broad match, Max Conversions or tROAS with a real target, consolidated budget, offline conversion import feeding actual closed revenue back to Google and spend all your time on offer, landing pages, and asset volume.

​Cold email - Precise list, short email, one specific offer, high volume, hygiene done once. Measure meetings held and pipeline, nothing above it.

​Cold calls - Good direct dials, clean caller ID, high volume of conversations, a 10-second opener you've said 3,000 times, one qualifying question, and a clear ask. Practice the first ten seconds and the close. Everything in between is improvised because it has to be.

​Posting on X - 1 real idea per post, in your own voice, at volume. Post from work you're actually doing your numbers, your failures, your take that would get you argued with. Link when there's something to link to.

The thing all eight have in common is the 'feeling of work' trap. At some point, tuning the machine becomes a substitute for having something worth putting through it. In the words of one of the greatest marketing slogans of all-time: just do it.

How to Become a Robotics Engineer in 6 Months

​Ronin put out a 6-month roadmap to become a robotics engineer and he argues that robotics is the least crowded high-value skill in tech. He says, "You cannot build a moat out of something a subagent does at 3am for pennies."

Robotics is different for reasons that are structural. The models already work, but what's missing is someone who can put them in a body. You cannot scrape the physical world, so someone has to move a real machine to make the data. Atoms are much harder to clone than bits.

He pulled live job listings from Figure and Skild and the same four requirements show up in every one (a degree is not on the list):

1. C++ and Python both
2. Real hardware experience
3. Depth in one area with literacy across the rest
4. Debugging listed as its own skill

The guide is almost 12,000 words with over 100 resources and 27 projects, and every month ends with something you actually have to build.

Figuring Out Your Next Career Move

​Lenny Rachitsky published a hands-on guide to figuring out your next career move, built from interviews with over 1,000 professionals changing jobs and hundreds of one-on-one coaching sessions. The goal is to help you take a step back and connect with what you actually want from your work and life.

The biggest issue with career development is that it is not always a conscious choice. People climb the default corporate ladder or get disrupted by events they cannot control.

Both are true. A guide will not prevent the events beyond your control, but it makes you far more likely to know what to do after them.

P.S Can you please respond to this email and bring it into your 'primary' inbox? You can say 'Hi!', tell us the last book you read recently, or what your favorite resource was from above.
​
It would also help if you add us as a contact on your mailing list, with a visual tutorial below:

We appreciate any feedback you are able to provide here. What do you want more or less of? Other suggestions? Feel free to reach out to us on Instagram and give us a follow there, tag your friends on our posts, and please forward this newsletter along to anyone else who would enjoy it.

Disclaimer: Becket U is an Amazon Associate and purchases through Amazon links may earn a small affiliate commission, but the price is the same for you. We only recommend books we love and think you would love, too.

Always wishing you the best,

J.B.