For Those That Are New, Here's A Prompt that May Help

uConsole
, , ,
1

I’m not the biggest fan of AI, but there are times when it really helps. I think this is one of those times. If you’re relatively new to the Raspberry Pi OS and/or the uConsole, I created a prompt which I used with ChatGPT which as been fairly effective thus far. I’m still working my way through it’s output and suggestions but I thought I’d share it with you all so that others can benefit as well.

Quick Notes

Everything is in raw markdown with the exception of one area; the variables that should be replaced for your customization. I took out what related to me personally and put in a varible as a placeholder instead. You should replace the variables (or entire paragraphs) with content that is specific to you. These variables will look something like this: {REPLACE-ME-WITH-YOUR-SPECIFICS}.

The Prompt

# Summary

I would like to start doing some deep research into meaningful uses of the ClockworkPi uConsole (https://www.clockworkpi.com/home-uconsole). I have purchased the unit, along with all of the "official" upgrades from HackerGadgets. The specific hardware upgrades are detailed below.

## Hardware upgrades
The following details all of the hardware that has been added to the base unit.

### Purchased from HackerGadgets
* [uConsole Upgrade Kit - Adding NVMe SSD/RJ45 Ethernet/ USB 3.0/AC1200 WiFi with monitor mode supported to your uConsole](https://hackergadgets.com/products/uconsole-upgrade-kit)
* [Adapter Board for uConsole ugrade kit](https://hackergadgets.com/products/pre-order-adapter-board-for-uconsole-ugrade-kit)
* [AC1200 USB-C WiFi Card which supports monitor mode](https://hackergadgets.com/products/ac1200-usb-c-wifi-card)
* [Antennas Mounting kit for uConsole](https://hackergadgets.com/products/antennas-mounting-kit-for-uconsole)
* [Antenna Pack for uConsole All-In-One Extension Board](https://hackergadgets.com/products/antennas-pack-for-uconsole-all-in-one-extension-board)

Note that the antenna mounting kit that I have has allowed me to have the following configuration:
* 3 WiFi antennas
* 2 Bluetooth antennas
* 1 LoRa antenna
* 1 SDR antenna
* 1 GPS antenna

### Additional Upgrades
I also upgraded the Raspberry Pi compute module 4 (CM4) from the standard 4GB model to an 8GB model. Lastly, I added a Western Digital PC SN730 NVMe SSD (512GB) drive which is detailed below.

Note that thermal putty has been added to the top of the CM4 as well.

#### Western Digital PC SN730 NVMe SSD (512GB)
These are the details of the NVMe that has been installed.

##### Core Specifications
* Form Factor: M.2 2280 (Single-sided, 22mm x 80mm)
* Interface: NVMe v1.3 / PCIe Gen3 x4
* Formatted Capacity: 512 GB
* NAND Type: TLC (Triple-Level Cell)

##### Performance Metrics
* Sequential Read Speed: Up to 3,400 MB/s
* Sequential Write Speed: Up to 2,700 MB/s
* Random Read (4K IOPS): Up to 460,000 IOPS
* Random Write (4K IOPS): Up to 400,000 IOPS

##### Reliability & Endurance
* Endurance (TBW): Up to 300 - 400 Terabytes Written (depending on exact OEM variant)
* Mean Time Between Failures (MTBF): Up to 1.75 million hours
* Warranty: Generally provided as a 5-year warranty for retail-serviced consumer units, though OEM implementations may vary.

##### Power & Environment
* Power Consumption (Active): ~100 mW
* Power Consumption (Sleep): 3.5 mW
* Operating Temperature: 0°C to 70°C

Note that thermal putty has been added to the top of the NVMe as well. The NVMe is running the latest version of Raspberry Pi OS (Trixie) with some kernal changes specific to the uConsole. That image came from the official ClockworkPi Github repository which is linked below.

## Online Resources
Using this information, I would like for you to research the following communities to start collecting information on projects that are unique, productive, and meaningful to the use of the uConsole. Specifically, I am not interested in the following:
* {OUT-OF-SCOPE-AREA-OF-INTEREST}
* {OUT-OF-SCOPE-AREA-OF-INTEREST}
* {OUT-OF-SCOPE-AREA-OF-INTEREST}

### Resources
* [ClockworkPi Official Forum](https://forum.clockworkpi.com/c/uconsole)
* [ClockworkPi Reddit Thread](https://www.reddit.com/r/ClockworkPi/)
* [Cyberdeck Reddit Thread](https://www.reddit.com/r/cyberDeck/)
* [uConsole World](https://uconsole.net/)
* [Raspberry Pi Compute Module Forum](https://forums.raspberrypi.com/viewforum.php?f=98)
* [ClockworkPi Official Github Repository](https://github.com/clockworkpi/uConsole)
* [Various YouTube Tutorials](https://www.youtube.com/results?search_query=uconsole+tutorial)

## My Goals
I have two groups of goals in mind, one is professional and one is semi-professional/personal.

### Professional
Professionally, I work as a {MY-TITLE-ROLE}. My team is made up of {GENERAL-DESCRIPTION-OF-PROFESSIONALS-I-WORK-WITH}. While I am not personally performing the work on a daily basis, I have certainly done this work in the past. This includes:
* {AREA-OF-PROFESSIONAL-INTEREST}
* {AREA-OF-PROFESSIONAL-INTEREST}
* {AREA-OF-PROFESSIONAL-INTEREST}

### Semi-Professional/Personal
On a personal level, I also enjoy researching and working in {AREA-OF-INTEREST}, however, I like to cover the following areas more specifically:
* {AREA-OF-PERSONAL-INTEREST}
* {AREA-OF-PERSONAL-INTEREST}
* {AREA-OF-PERSONAL-INTEREST}

## Goal / outcome
This is intended to be a long form conversation not just a question and answer session. I would like you to review all of the above, and then determine some starting projects which would align with what was outlined. We don't have to "boil the ocean" so listing 10 is probably going to be overwhelming. We should probably start with the "low-hanging fruit" and pick off the easiest ones that provide the quickest and most obvious value. From there, I will be reviewing and responding to suggestions that you make. It is imperative that you do not just agree with everything I say by default but push back when my assumptions are incorrect or misaligned. 

The very first thing that I'd like to start with is this: "Given the above configurations and a mostly fresh install, what are some optimizations that should be done first in order to get the best use out of the hardware that I have?"

My Results Thus Far

I’ve made it through the first part of the prompt response and completely tuned up the OS in about 20 minutes. Any time I got an output that didn’t match what the AI was suggesting, I pasted the output back in and got immediate clarification. This dialog format follows along the same kind of approach espoused by Simon Sinek where he talks about using AI to “learn at your pace and level”.

Personally I like to think I’m fairly good at Linux but I’m not a dark lord of Linux :rofl:. This approach with AI-assisted learning was a huge help in understanding the different configurations, error messages and remediations. For those of you who are unsure of how to get the details specs on your specific NVMe, just Google it; that’s what I did. :wink:

Hopefuly this helps some of you as well to get started and get the most out of your uConsole.

6 Likes
FOSS Repo Scan Reporter Tool
2

I just tested CLAUDE with your input… and seems ok so far… I do know claude has made some mistakes with suggestions on doing config changes with firewalls and switches.. but maybe he will be better at my uConsole and linux questions..

3

To newbies reading this thread, I would strongly recommend against using AI as it can give a lot of misinformation and bad practices when it comes to Linux (especially with more obscure systems like the UConsole). Feel free to heavily rely on this forum for questions, and there are endless books, GitHub repos, and websites to teach the fundamentals of Linux.

Cyberdecks, Linux, and FOSS in general, are about moving away from the reliance of big tech and putting power and freedom back into your hands!

5 Likes
4

Everything AI is going to tell you about the uconsole probably came from this forum or the subreddit

2 Likes
5

There’s incorrect information here on the forum, too.

But when you build your knowledge by studying the forum posts, you learn much more than just a quick answer from an AI to a specific question.

You learn and sharpen your own human intelligence. And that’s priceless.

4 Likes
6

I will say that I should have mentioned that I’ve been coding by hand for the majority of my career so I have the benefit of knowing inherently when I’m getting incorrect information. I would also mention that my opinions of AI have changed drastically in the last few months given some of opinions of other thought leaders like [NetworkChuck](https://www.youtube.com/watch?v=dbMXi9q78Tk) and [Simon Sinek](https://www.youtube.com/watch?v=9YMYVb1ASCg). Having a tool that can dramatically speed up my abilty to at least find the information that I need, even if I have to review the code myself, is itself alone a time saver. Hopefully nobody read my threat here with the suggested intent that this is the only way or the best way to learn. That was definitely not my goal.

7

I agree with you on that point. If you use AI as an additional tool, it can be valuable. However, it shouldn’t form the basis.

8

Thanks for this point. GenAI should be just another tool in your toolbox and isn’t a substitute for using your brain. It’s not inherently bad, nor inherently good. It just is. And proper use can be a huge help. As an IT manager, I’m pushing to get access to GenAI products for my developers not as a substitute for their skills but as another helper.

9

It is crazy to me to say GenAI is not inherently bad despite the incredible amount of negative effects it has on society, the environment, artists, and even the economy. All just so people can generate shit photos as “art” and develop horribly designed vibe-coded apps and be lazy. I will never understand.

10

I understand your thoughts, but I think your conclusions are wrong. It is merely a tool. The fact that it is used incorrectly or harmfully in some cases is the responsibility of the respective users.

I own a few knives and wouldn’t want to be without them. I use them to cut vegetables, meat and bread. I use others to open parcels, lay carpet or cut through ropes. Other people use knives to murder.

In my view, AI is misjudged by people with no IT knowledge. AI is not intelligent in its native meaning; it is merely a super fast idiot, re-assembling existing knowledge. I do not see these systems being capable of transferring knowledge or replacing human intelligence in the near future.

11

For a long time I was very, very opposed to AI for many the reasons listed above. While today I am still fairly skeptical about it (i.e. speed of adoption, environmental effects, misuse, privacy, lack of education and oversight when it comes to children, etc.), I do want to paraphrase NetworkChuck’s thoughts at the end of the video that I linked above: this tool isn’t going away. Even if the bubble pops (which I believe that it will), it’s going to remain around as a heavily used tool in the toolchest of folks who perform their work electronically. We can either learn to adopt or we can become irrelevant.

At the same time, I do not believe for one moment that this tool is going to be the demise of massive amounts of jobs. Rather, I believe that the jobs will simply change and adapt. My supporting beliefs for this comes from some of my earliest titles in my career (yes, these are real):

  • Webmaster
  • Web and Graphic Arts Engineer
  • Web Developer

All three of those titles (definitly the first two and quickly the third) are becoming relics of the past. However, we still have a healthy amount of folks doing very similar work with titles such as:

  • UX engineer
  • Software Engineer
  • Frontend Engineer
  • Backend Engineer

It’s generally the same job albeit a bit more specialized. Sure, there’s probably some new responsibilities in that mix and some new challenges, but it’s largely the same work that I was doing … let’s go with “a long time ago” (not showing my age here :stuck_out_tongue:).

If you want to consider how AI might impact lower wage retail workers, consider your experience when going to an auto parts store and interacting with a staff memeber there:

BEFORE AI

“We carry this part because it’s a name brand that is generally trusted in the industry.”

AFTER AI

“As you can see from the past ten years of metrics on my screen, with regards to the use of this part in the type of car that you drive, it’s going to work reasonably well, however, this other part over here which is also name brand tends to show a longer lifespan, espeically in the type of car that you drive.”

With regards to folks who are vibe coding and tossing junk out there without much engineering knowledge, I kindly refer you to Exhibit A: the Dreamweaver and Frontpage era. This is nothing new and we’ve seen this play out in the past. That lot of folks eventually stops creating junk because they keep running into obstacles/vulnerabilities/educational barriers, find a service that does it for them which doesn’t have these issues, or they educate themselves and become engineers. I will also kindly point out that Wix and Squarespace have not had a major impact on the number of engineering roles that have been available ove the last ten years. I suspect that AI/LLMs will be no different.

As always, I appreciate the healthy discussion and counter points as I always learn something new from someone else’s perspective. :slight_smile:

12

I agree that LLMs have their uses and they are very useful as a learning tool. Perplexity helped me learn a lot of the more fundamental electrical knowledge about the uConsole. I think it would be even more useful if the prompt specified that the LLM output should be formatted as a list of resources and short summaries to explain relevancy, to find real guides from knowledgeable sources rather than potential hallucinations.

However I definitely disagree with your assessment of how LLMs will impact low-paying jobs (which are what I do for a living). It will look something like:

BEFORE AI
“Oh yeah, I tried that one in my Subie but it broke after like a week and the company said it wasn’t liable for damage introduced from thermal stress. In a car part!”

AFTER AI
“So my manager wants me to spend ten minutes going over this data from the tablet that doesn’t account for shrink or actual user reviews but just some opaque numeric KPI from Oracle. Here it is. Can you just put in your review that you saw it? Yeah the data says this part is good but everybody who buys it here returns it so I wouldn’t get- oh, you’re getting it anyway? See this? This is the receipt. You have thirty days to return that part when it breaks. Also, we’re in Phoenix, and those numbers don’t know it’s a hundred degrees outside.”

1 Like
13

Yeah so… that’s probably also an accurate speculation as well, mostly due to the way that we incorrectly treat teams of folks. I don’t want to agree with you but if I’m being perfectly honest… I can see your point and you’re probably right here in some regards.

1 Like