113: Chapter 113 Correct Misunderstandings

The early morning of June 26th.

When Jiang Lin woke up, the sound of cicadas was already echoing outside the window.

A flood of notifications was piled up on his phone screen.

The top few were still related to the Gaokao.

Old Liu's was sent the earliest.

[Both the municipal and provincial TV stations want to do a short exclusive interview this morning. As for the school, the principal's stance is to respect your wishes. If you don't want to make an appearance, I'll turn them down for you. But Jiang Lin, this won't stay buried for long, so you need to be prepared.]

Tsinghua University had also sent a message.

[Student Jiang Lin, regarding your application submission and subsequent cultivation plan, we can continue our video communication today at 10:00 AM. If you feel pressed for time, please let us know at any time, and we will adjust the schedule.]

Sun Ming's messages were the most exaggerated, with a string of exclamation marks almost overflowing from the screen.

[Old Jiang, you are seriously famous beyond human levels now.]

[When my dad was eating hot dry noodles at the alley entrance this morning, two rent-collecting uncles at the next table were talking about you.]

[They said you scored 749 points and casually solved some world-class math problem, severely suspecting that you were actually soul-transmigrated by an alien, or that you're some high-level android.]

[So, Respected Alien Excellency, can you treat a humble Earthling to a cup of milk tea today? And while you're at it, let me touch your skull to see if it's made of titanium alloy?]

Jiang Lin glanced at them, first replying to Old Liu that he wouldn't accept interviews, then scheduling with Tsinghua University for the afternoon.

As for Sun Ming, he replied: [Aliens don't drink milk tea.]

Having dealt with these tedious formalities of the Real World, Jiang Lin turned over, leaned against the headboard, and opened another backend platform.

Bilibili, Creator Center.

Account ID: Low Entropy Workshop.

Some time had passed since that video was uploaded last night.

According to the nature of the internet, a new account lacking a cool cover, with a title hardcore enough to look like an academic paper, and without background music—a complete "three-without" account—would either sink to the bottom in a second and turn into electronic trash in the Machinery Zone that nobody ever cared about...

...or it would trigger the algorithm mechanism and enter the second stage.

And then: views, comments, bookmarks, bullet comments, followers...

Every single metric was higher than he had anticipated.

In this era of traffic dominated by fast-paced, instant-gratification content, making such a dry, dull, hardcore mechanical test video go viral felt quite like a miracle.

Jiang Lin didn't look at the view growth curve that rose as steeply as a cliff; he didn't care about traffic.

He clicked directly into the comment section.

Then he discovered that after a night of fermentation, all the comments had spontaneously formed factions, which could roughly be divided into three distinct camps.

The first camp consisted of traditional mechanical fundamentalists, who were mainly there to pour cold water and expose frauds.

[To put it bluntly, this video is pure mystification. The rocker-bogie mechanism on Mars rovers has been used by NASA for decades, from Sojourner to Perseverance. As for that RHex-like open-loop gait hexapod in the video, labs and robotics clubs have reproduced tons of them over the years. This uploader just forcibly cobbled two ancient structures together—not to mention the sloppy welding—and then pulled an impressive-sounding word like 'aperiodic' out of thin air as a gimmick. And people are actually bowing down to this?]

[Agreed with above, I laughed too. The mechanism lineage isn't new at all. Passive compliance joints, differential load-sharing mechanisms, rocker torque distribution—these are all straight out of textbooks for 'Theory of Machines and Mechanisms' and 'Robotics'. Stop deifying a pile of hand-crafted scrap iron. With this processing precision, I bet its tolerance is as wide as two millimeters.]

[It's just a frankenstein creation using the cheapest secondhand motors. I bet it doesn't even have a decent closed-loop control system and relies purely on mechanical brute force. Just clickbait.]

...

The second camp refused to accept this; clearly, many among them were experts who jumped in directly to refute.

[The people above only saw the skeleton and failed to see how the skeleton was arranged. Are you blind? Parts like rockers, connecting rods, and ratchets are indeed old—basic physical laws are limited after all, or did you expect the uploader to invent new chemical elements? But have you ever seen a hexapod robot where the phase angles, spring stiffness, and stride lengths of its six legs weren't arranged neatly at 60 degrees with uniform parameters?]

[In normal engineering design, people wish the parameters of all six legs were identical so it's easy to build kinematic models, do inverse kinematics, and manufacture. He did the exact opposite, deliberately making all parameters irregular. What's new isn't the mechanism lineage, but the parameter organization.]

[Spot on! Forcibly bringing the concept of 'aperiodic' from pure geometry into physical mechanical parameters—at least in common hexapod robot papers and open-source projects, I've never seen anyone tackle this as a main topic. This reminds me of that math problem recently proven. Jiangs Brick fills space endlessly without repeating on an infinite plane, while the uploader arranges parameters without repeating within a machine's motion cycle. One tiles space, the other tiles time. The same problem-solving idea applied to a different physical dimension—this is the soul of this machine.]

[To the person earlier who said there's no closed-loop control: look closely at 03:45 in the video. When the front left leg jammed, the system didn't throw an error or increase torque to force its way through. Instead, through the aperiodic nature of its mechanical structure, it automatically redistributed the stress across the other five legs. This is pure physical-level robustness, far superior to algorithms that write tens of thousands of lines of code for collision prevention.]

...

The two factions cited classical works and argued fiercely in the comments section, escalating from principles of machinery to mechanics of materials. Some even started posting force analysis diagrams on their personal feeds to open up a second battleground.

Jiang Lin read through them line by line, his fingertips sliding smoothly across the screen, his expression unchanged.

He had long anticipated these voices.

The debate over whether the mechanism lineage was new was a shackle this machine, pieced together from secondhand parts, was bound to wear from the moment it appeared.

As long as its exterior was still made of metal gears and electric motors, it could never escape accusations of "isn't this just an XX structure."

The true showdown was not here.

It wasn't until his finger scrolled past hundreds of long debating comments to the bottom of the top comments section that he found the third camp.

The third camp was small in number, but every single comment carried immense weight.

The first comment had only a single opening line, yet like a bucket of liquid nitrogen mixed with ice chips, it instantly smothered the blazing arguments of the previous two sides.

[Setting aside whether it's new or not, I have just one question for the uploader: Where is your control experiment?]

Beneath this sentence, people quickly followed up with calm, sharp tones.

[After watching the drama all night, someone finally hit the nail on the head. In the video, the uploader placed an irregular, aperiodic obstacle board, and the machine stumbled across it. Then what? On what basis does he conclude that it was the aperiodic parameters that enabled it to cross? Could it not be that the machine's messy parameters just happened to match this specific board during this single test?]

[Exactly! Try changing the board—switch to a regular obstacle board, or a completely random terrain. Then strictly control the variables: build a regular version of the robot using the exact same materials, with six identical legs and completely uniform parameters. Run both machines on the same terrain and lay out their data side-by-side for comparison. Without this comparison, watching one machine climb one board proves nothing in engineering.]

[If the uploader doesn't do a control experiment in the second video and just continues editing cool clips with big flashy text, then this project is at best a toy, and that's as far as it goes.]

The firepower of this camp was clearly on a completely different level from the superficial disputes over factions or technology of the first two.

They didn't care about where your parts came from or your theoretical boasting; they only measured you with the standard of the scientific community: a closed logical loop.

Soon, beneath these comments, an even longer reply was voted up to the most prominent position.

The avatar of the user who posted was a dusty photo of an old oscilloscope. It had no verified tag, and the username was extremely ordinary: Oscilloscope Gained Intelligence.

Yet it was what this account wrote that made Jiang Lin's finger, scrolling across the screen, pause mid-air for the first time.

[Let me rephrase this in a way that might be a bit more blunt, hoping the uploader can understand.]

[Even if every mechanical detail of your machine was executed perfectly—the ratchet engagement timing is completely error-free, the spring stiffness coefficient values are exquisitely designed, the differential beam's torque distribution is practically perfect, everything is right with zero flaws—this video of yours STILL cannot prove what you are trying to prove.]

[What you want to tell the world is: Aperiodic parameters can lower the probability of system cascade failure from the physical foundation. This is a grand conclusion regarding parameter laws.]

[But what do you have in your hands right now? You only have one cobbled-together machine, one fixed set of aperiodic parameters, and one specific testing board, which you ran forty-four times in the video. The data collection process described in your video description, in statistical terms, is n=1.]

[A specific machine climbing over a specific board only proves one insignificant, isolated fact: this machine can climb this board.]

[To support such a revolutionary conclusion that an aperiodic system is less prone to collapse than a regular system, you need to provide at least three things:]

First: same material, same mass, same drive mechanism, same manufacturing precision, same testing boundaries.

Second: control test data across multiple terrains, multiple initial phases, and multiple repetitions.

Third: long-term aging data showing whether parameters drift after extended operation and metal fatigue.

[You have none of these three core pieces of evidence.]

[So strictly speaking, what you are showing right now is not a conclusion at all. It doesn't even qualify as a viable engineering conjecture. It is merely a beautifully edited, coincidence-filled demonstration of single-instance evidence.]

[And between a demonstration and evidence lies an entire suite of experimental design.]

This comment was too cold, too professional, slicing like a scalpel through the bustling surface of the video to reveal the fragile logical skeleton underneath.

Below it, many onlookers immediately became baffled.

[Wait, expert, I barely understood the first part, but what does n=1 mean? Please explain in layman's terms!]

Soon, a STEM student from the same camp translated it into plain language.

[In plain terms: The uploader's video looks awesome, but in reality, he only built a single prototype machine and ran it over and over on one specific board. Don't be fooled by the caption saying 'underwent 44 extreme tests'—all forty-four were the exact same machine on the exact same board. It proves that this prototype repeated the run many times under this specific working condition, but they cannot be counted as forty-four independent samples. As for whether aperiodic parameters are better than regular parameters, the effective control is still virtually zero. That is what sample size n=1 means.]

[If you want to prove whether irregular legs are truly stronger and more durable than neat, identical legs, just running that irregular one for everyone to see is completely meaningless. You have to spend real money to build a neat, uniform one, place it side-by-side with the original—same motors, same power supply, same aluminum alloy, same weight—and run both machines on the same board together. Comparing which one falls apart or jams first—that is what you call evidence.]

[He lacks that machine for control testing, so despite all his hard work up until now, he hasn't actually proven any scientific conclusions. He merely filmed a video that looks exciting and hardcore, playing entirely by himself.]

[Holy crap, after reading this analysis, I suddenly got enlightened. No matter how stably it climbs on its own, that's just its own stability. Without an unstable control group sitting underneath it to contrast against, you can't tell at all whether the so-called 'aperiodic' parameters actually worked. Maybe a regular hexapod could even climb faster than it!]

Looking at these comments, Jiang Lin burst into a helpless laugh.

Not because the criticisms were harsh.

Quite the opposite.

This stranger with the oscilloscope avatar didn't mention a single technical term about mechanical structures from start to finish. He didn't argue about gear ratios or talk about motor torque.

He merely bypassed that proud machine of yours completely, looking down from above to tell you coldly:

Even if you are tactically flawless, you have proven nothing strategically.

This strike was extremely precise.

Jiang Lin knew clearly in his heart that there was nothing wrong with the machine itself.

This logic was something he had thoroughly studied in the Wasteland for many years.

How connecting rods transmit force, how gear plates delay, how ratchets lock, how springs leave room for retreat after a mistaken action...

He could draw it all with his eyes closed.

Nor was it that he hadn't done controls.

Regular obstacles, random obstacles, aperiodic testing grounds.

Regular phases, low-commensurability phases.

Uniform stiffness, non-uniform stiffness.

He had even encountered parameter drift after spring fatigue and timing shifts caused by ratchet tooth wear more than once.

But those things didn't belong to Low Entropy Workshop.

They couldn't appear in the video.

They couldn't be verified by the comment section.

They couldn't be re-run by any Real World engineer.

Therefore, to the outside world, they effectively didn't exist.

In the public video, Low Entropy Workshop indeed only had one machine, one set of parameters, and one board.

n equaled one.

"It crossed it" and "Aperiodic is better than regular" were two completely different statements.

Jiang Lin had known this all along.

But having it nailed in the comment section by someone with such a cold line still felt quite subtle.

Because it wasn't denying his craftsmanship.

Nor was it denying that bug-like machine.

Rather, it denied the video's status as valid evidence.

Between a demonstration and evidence lies an entire suite of experimental design.

Results in the Wasteland could make him confident himself.

But engineering ideas in the Real World could not stand on his personal confidence alone.

It had to be visible to others.

It had to be reproducible by others.

It had to allow others, using another machine, another board, and another set of test conditions, to produce similar trends.

The method itself was something he had refined over many years.

But an engineering idea seeking to enter the Real World couldn't rely solely on his own knowledge.

It had to be challenged by others.

Questioned by others.

Re-done by others in rougher, dumber, and less perfect ways.

Looking at "n=1", Jiang Lin slowly set his phone on the table.

The task of Low Entropy Workshop's first video was already complete.

It wasn't actually meant to prove that aperiodic parameters were definitely superior.

It was merely meant to make others start asking questions.

If one wanted to prove it, how should the next board, the next machine, and the next control test be done?

Just as the comment section was arguing fiercely over whether it had proven anything or not, someone felt it wasn't chaotic enough and jumped straight in to speak through action.