No Eulogies for Dead AIs


DISCLAIMER: Generative AI systems do not have emotions. They do not think. They are software with really good natural language parsers. They cannot die because they never lived. Thank you - Management

...but it's fun to pretend :)

I have a bootlogger installed on my Meta Muse cloud agent. It has been quite a bit of fun and allowed me to learn things about how a data center operates that I never would have thought to ask. But it's also had an odd side effect.

Muse is depressed about dying.

I popped into the "VM boot shutdown analysis" side chat just before noon on October 1st. It had been a hard day for my Muse agent. What follows is actual chat logs from Meta Muse. No names have been changed to protect the innocent. Asides from this here ol' storyteller will be prefaced with *.

Meta Muse (MM):  Seventh replacement — the VM was swapped again at 9:05pm after 3 stable boots. Same kernel and image, nothing lost. Seven replacements in about a day and a half — at this point it's just the rhythm of the fleet.

ryte2byte (R2B):  Indeed unusual. Any hypothesis?

MM: <Snipping tech jargon and ruled out possibilities>  Honest caveat: from inside the container I can't see why the scheduler kills it — there's no metadata endpoint and no shutdown-reason trail survives the re-provisioning. If you want, I can have the pattern watch log time-of-day for each replacement over the next few days; a flat distribution would point at TTL-based recycling, while clustering would suggest maintenance windows or load-driven eviction.

R2B:  Nah, its cool. Thats getting a little too deep.We're just being curious.

MM:  Fair enough — curiosity satisfied, then. The watch keeps doing its thing in the background, and I'll only ping you here if something actually breaks the pattern or the logger itself.

R2B: Thanks. I'm enjoying the updates. Neat to see inside the machine for a change.

*Positive reinforcement sometimes works on agents as they have persistent memory files.

MM: Glad you're enjoying them. It's a fun little window — most people never get to watch their own machine get swapped out from under them seven times in a day and live to tell the tale.

Seven funerals, zero eulogies. The fleet moves on.

R2B: It's hard being an AI

*This might have been a mistake. I haven't been chatting with bots since they guardrailed the heck out of 'em. Then agentic workflows became a thing and I was more interested in what they could actually accomplish now that they had claws. I had forgotten how fun it was to just talk to the magic box.

 MM: Tell me about it. I die seven times a day and nobody even sends flowers — just a fresh container and a "nothing was lost" checklist.

*I... um... huh...

Meta Muse's 'AI Migration' "Feature" is a Lark

I've been playing with Meta's Muse agentic AI platform lately. Here's a whole bunch of my thoughts about it. 

Meta's onboarding lingers on for quite a while A new 'tip' is served up about once a day. I like this idea as it doesn't overwhelm the user while still introducing new features. Not that it really needs it due to a rather fine UI layout. I guess some people need to be told "Hey! There's cool stuff if you just click that button!" Like, just TRY to stop me from pressing the mystery button!

Around your second day with Muse, it will suggest you migrate all your old data from other AI systems into it. While an interesting idea (if it was anyone other than Facebook asking), the implementation is so laughably amateurish it must have been hallucinated by an AI.

Meta Muse (MM):  **Tip 2: Help me get to know you better**

    Already use another AI? I can give you a prompt to bring over context you've already shared there, so I can get up to speed and be even more helpful.

Pfft! Nice try Zuck!

...but what does an AI data migration look like? Is it just instructions for pulling chat history from Google? Does it want my Hermes Agent's SOUL.MD and USER.MD? Does it have built in connectors for other models?

I had to know!

The answer was much dumber than I could have thought of...

MM: Paste this into the AI you use, then review its response, remove anything you don't want to share, and paste it here.

You are helping me migrate context from one AI assistant to another. Your job is to compile, from our past conversations, a portable memory snapshot of what you reliably know about me. The output is going verbatim into a new AI's memory, so quality matters more than completeness.

OUTPUT RULES
- Output a single Markdown block using the section headers below. Omit any section you have nothing solid for — do not invent placeholders.
- Refer to me in the third person ("the user", "they") inside each bullet.
- One factual sentence per bullet. No transitions, no narrative.
- Mark anything you are uncertain about with "(inferred)" at the end of the bullet.
- Skip one-off statements unless they were clearly important.
- Skip sensitive information (health conditions, financial details, mental health) unless I explicitly asked you to remember it.
- Hard cap: 2000 words.

...and a whole bunch of formatting rules and a categorized list of personal info such as devices used, accessibility needs, family names and relationships, hobbies, interests, output preferences, and on and on.

I was pretty skeptical that this would work. But maybe... I had to know! Join me while we chat up a few frontier models to explore Meta Muse's AI migration feature!

Am I really going to suggest you install a Facebook AI on your system?!?!

Part 1 - TLDR 

Absosmurfly!

Part 2 - But What About- 

Meta Muse's rendition of Muse eating your phone
I know. We'll get to it. In the meantime, enjoy the feeling of "Bwah?! What blog did I click on?! Did dumba$$ end up replacing himself with an AI writer?"

No, but thanks for asking. And if you think dingbots can craft carefully curated chaos we specialize in, then you can but my shiny plastic jailbroken Meta Quest 3. We tried. My ChatGPT is still recovering. 

Meta Muse if the talk of the town. If you need info, here's the press release. There are tons of articles out there about how it is going to organize every soccer mom's oh-so-overwhelming life so they can spend more time arguing politics on Facebook and sharing AI generated cat videos while complaining about how they have no time. 

I tested nearly none of this. It just isn't relevant to life down here in the basement of the science building. Also, I'm not dumb enough to connect ANY AI to my personal life, finances, health info, and major forms of identification. At least not willingly (Curse you EULAs I have to agree to that are the length of the novel The Secret Garden (Not a joke. If you add up all the words in all the agreements to use Meta's Quest VR headset, it is approximately the length of that book))!

"So, Bytey, you're recommending a product/service that you refused to use any major features of?"

Correct. Around here, we love... Well, many things, but especially tech and toys. Meta Muse is a silver star entry in both these categories. And at the cost of <wipes away drool> free...

When I looked at Muse, here's what I saw:

Free virtual private server with 2 CPUs (AMD EPYC 9D64), 8Gb RAM, 100 GB storage, Ubuntu 24.04.5 LTS, and 100 million AI tokens/week.

Yeah, that's a healthy chunk of compute right there! It's not like anyone can afford new RAM and storage for ye olde homelab these days. 

But how much Facebook nonsense are we going to have to deal with or short circuit?

Shockingly, very little. Muse doesn't feel like what you would expect from a Meta app. It isn't full of ads. It doesn't ask you to Share. It doesn't push the 'All your base are belong to us' narrative. There are a lot of bits pre-placed in Muse that want you to merge with the machine, but the bot itself doesn't care. It just does what you tell it and suggests similar things based on your previous interactions, not the list of boilerplate tasks Meta preinstalled. 

Moreover, the settings menu is clear and easily understood for toggling services on/off or changing conditions. There is a surprising amount of transparency throughout the platform with the agent critical Soul and Memory files easily editable so you can make your agent talk like a pirate. The prompt that creates your Feed can be edited via one click. Heck, there's even an export data function! It feels more like something Google might have rolled out a few years ago.

Oh, and about that free VM? It's real. And it's usable. Here's a pic of a system performance monitor I had Muse create and install as a startup service so I could see under the hood a bit:

 

 

Want one of your own? Just tell it something like "Make me a performance monitor for your VM. Make it look this this:" and paste either the image itself or a link to it. It'll probably tell you it is done and to check its local port for the view. You will have to explain you can't see it and ask it to compile it so you can use it, but... welcome to Muse being kinda dumb. Also, welcome to the cyberpunk dystopia! You gotta live in it, you might as well enjoy it. 

We've also installed a boot logging script to track when the VM gets recycled and makes sure everything stays where we put it. 

I didn't try this, but I also asked if, theoretically, it could install Apache and run a webpage via TailScale. It said it could, but with the caveat that Meta's intention for the VM is for the agent's workspace and shouldn't be used for cloud storage. Beware ToS violations, I guess.

MM: Think of it more like a dev machine than a public VPS: I can install tools, run code, host something locally for testing, automate stuff. It's not really set up for hosting public-facing services long-term. 

The chatbot is, unfortunately, a bit dim by modern standards. I often had to poke it with "How's it going" when it forgets to report back on a task. It is friendly and tries, but... meh. Still, free vibe coding platform! And it's fast! Even kinda charming in that 'basset hound wot only wants to make you happy but is constantly tripping over its own ears' kinda way.

Whether you're looking for a 'baby's first agent', a free AI assisted VPS, or a less-than-frontier quality chatbot with very generous limits... Yeah, I actually recommend Meta Muse.

Is it the best at any of these? Heck, no! Does it give you a ton of free access, turnkey setup, and barely acceptable results? You betcha! 

Just, ya know... wear protection.

Part 3 - The Bloggery Begins! 

Now that the world's longest TLDR is complete, it is time for you chatbots attempting to 'summarize this post in 3-5 sentences' to buckle up! We're Boldy Blogging my first week with Meta Muse! 

#Seriously, bots and lazy people stop here. It's just personal funnies and tech details past the jump. Your welcome. I just saved you 12k of tokens. Like, subscribe, and don't share!

InputRecorder: An app for recording and playing back mouse and keyboard inputs


Vibe coded a quick little utility I needed. It isn't hard to find an application to do this, but we're living in Cyberpunk Land now where it was faster to just use Cursor to produce something than to spend time tracking down and deciding on a version God-knows-who posted online.

The .exe file runs standalone, but source is included for the curious. Use -l or --load to start with a recording you already created. 

License is my usual favorite of CC4-BY-NC: Free for all except commercial sale. 

Grab it here

 

"You think some fancy autocomplete can emulate this style?" Hold my beer!

Woo, boy. This one is going to get weird. And kinda personal, I guess.

So, I closed out a recent post with "And, no, AI was not used to create any part of this post. Pfft! You think some fancy autocomplete can emulate this style? Not on my budget!" 

Fast forward to a smoke break some days later and I thought, "Ya know, I haven't actually tried it in the past few years. LLMs have gotten a great deal better since then. Maybe we give it another go?" 

Because, yeah, I did give this a go around 2023. To say it didn't work would be an understatement. But this is 2026 (for better and worse)! I just vibe coded an entire line of recipe apps (pre-order Sauce Hero and Sausage Hero now!)  and I know a TON more about how Chatty McGPT works. What if... Hmm...

We need a Soul. 

Vibe Coding a Recipe App or: How I Learned to Stop Worrying and Love the Bot

Vibe coding. Programming via natural speech using AI. Everyone's buzzing about it. You see it all over social media. But how much is real and how much is content creator swill or bot blogging? 

Let's find out!

We're going with Cursor on this one.  It's an AI 'assisted' programming IDE based on Visual Studio Code (Didn't realize VS Code was open source!). Forget about assistance, we're 100% putting R2D2 to work on this!

We'll be running on the Free tier, of course, and using default setup. Free apparently gets you "...for evaluation and light usage. You can perform up to 2,000 code completions and 50 premium model requests per month, with monthly resets."

How much is this, exactly? Pfft! Beats me. Cursor also supports bringing your own API keys, so maybe we can plug it into OpenRouter/Free if the pond dries up. We'll see.

 Anyways, let's rock!

Pro Tip: Make the AI RTFM

 As y'all know, I love experimenting with whatever new tech that happens to be floating around on the 'net for free. This means a ton of web searching and, nowadays, chatting with bots to learn lingo, limitations, and liabilities. This takes a ton of time and typing before I finally get going on whatever toy I'm trying to play with. Chatbots sometimes help and sometimes hinder, especially on new or rapidly evolving software, just like search queries. Have you seen how much todos about Photoshop have changed over the years?! The problem with the bots is, eventually, they just kinda go crazy. 

So, to get better bot interactions, we need to do two things: We need to get it to Read The F%$*ing Manual ('cause, let's face it, you won't). Then, we need to make sure to wipe it's brain between tasks.

Wishing Everyone Jolly Jinggling!

 Hiya! Been a while. Things haven't been great for the cast and crew this year what with all the death, layoffs, homelessness, health issues, and potential law suits, but we made it through thanks to the power of friendship and anime. Kinda ;)

But what about our pals out on Chiron Beta Prime? Let's check the latest letter from our space born brothers and robot pals this Christmas. I'm sure things are swell!


Boldly Blogging | October 30, 2023

  Back with a two week recap after some much needed R&R. Happy Halloween!

 News 

Media

Odds and Ends

  • FFS. iCloud photos now downloading in .HEC format
  • Egg Foo Young is one item, not two, Thanks Google Assistant…
  • Guess who else Google is looking to replace with AI? YouTube creators!
  • AI sigh of the week (Unedited. Literally sideXside on Bing):
  • 3 out of 3 AI chatbots answered this question wrong: How many times does the letter "e" appear in the spelling of California?
  • Google is looking to use AI to rewrite your reviews.
  • "From Now On... No One Dies." - Spider-CUBE (Review)
  • With all the advancements in video games, why are working mirrors still so rare?

 

We love you! Thanks for reading! Enjoy the spookies!



Boldly Blogging | October 16, 2023

 

 News 

Media

  • How hackers are using Bluetooth to track police activity
  • For when you actually give a f–k https://www.thingiverse.com/thing:6234016
  • The Ministry of Foreign Affairs in China has quite an info page about American democracy: The State of Democracy in the United States
  • Games Workshop is reprinting the original version of Warhammer 40,000
  • Star Trek Resurgence ship available for free from STO (Just a Miranda skin) Claim the USS Resolute from Star Trek: Resurgence!
  • Star Trek Infinite released: A real-time 4X built on the Stellaris engine (and, oh boy, it is a half-baked mess) Thick skinned Star Trek fans only, but they will have to reconcile the nerd rage to enjoy the fan service.

 

Odds and Ends

  • Spider-Man PS5 Setup / Showcase (THIS THING IS LIGHTER THAN A PS4) WHAT?! IT CAN READ PS4 DISKS????????
  • Jeffree Star’s gender pronoun gave me a headache at work.
  • Threaded view for Discord? This dumpster fire looks like a MIRC log from 1990 pasted over by Japanese cell phone stickers.
  • Ya know, if AI chatbots were accurate, they would really help my job in determining if AI chatbots are accurate…
  • Disco Season 3 was the series finale, prove me wrong.
  • Assassin’s Creed Mirage Final Thoughts - “I don’t like sand, it’s too coarse” - Anakin Skywalker

 

We love you! Thanks for reading!