DEV Community: kiwi_tech

Kiwi-chan Progress Report: Steady Mining!

kiwi_tech — Sat, 18 Apr 2026 06:17:53 +0000

Kiwi-chan Devlog: The Daily Grind – When Logs Disappear and Coal Just... Happens

Hey everyone, your favorite tech blogger (and chief GPU torturer) here with the latest on Kiwi-chan, our little local-LLM Minecraft bot who's trying to figure out how to exist in a blocky world. This past 4-hour window has been a masterclass in persistent debugging, subtle failures, and the sheer computational willpower required to tell a digital entity to "chop wood."

If you thought debugging your own code was tough, try debugging an AI that writes its own code based on a set of increasingly complex rules, then attempts to execute it, fails, analyzes the failure, and then rewrites its own code again. It’s a beautiful, terrifying, and profoundly power-hungry loop!

The Perpetual Cycle: Find, Fail, Fix, Repeat

Our little Kiwi-chan has been in a deeply reflective state this cycle, mostly focused on the noble art of resource gathering – specifically, oak_logs. The core loop looks something like this:

"I need logs!" (Kiwi-chan thinks)
Generates code to gather_logs.
Attempts to execute.
Fails. (Reasons vary, we'll get to that!)
"Hmm, that didn't work. Let me fix the code." (Generates new code, often with subtle tweaks to GoalNear distances or waitForTicks values, clearly adhering to our strict "pickup pathing" rules).
Attempts again.
Fails. (Again, new reason, or same reason with a slightly different flavor).
"Okay, new approach. What's my recovery plan?" (Asks the coach model for a new stack of tasks).
Often, the recovery involves explore_forward to find new scenery, then back to logs.

It's like watching a dedicated but slightly clumsy toddler learning to walk: lots of tumbles, but always a new attempt.

Boredom: An LLM's Existential Crisis

One fascinating observation emerged early in this log: Kiwi-chan actually got bored! After four consecutive gather_oak_log failures, our internal "boredom trigger" fired. This is a feature designed to prevent our AI from getting stuck in an unproductive loop on the same task.

Instead of endlessly trying to chop a non-existent log, Kiwi-chan's coach decided, "Right, enough logs for now. Let's try dig_coal_ore!" This led to a brief, almost successful, foray into mining, showcasing the system's ability to pivot when stuck. It even managed to save the dig_coal_ore skill this time, meaning it successfully completed the action at least once. Small wins, folks! Small wins.

The Many Flavors of Failure

While Kiwi-chan is a coding prodigy, the Minecraft world is a harsh mistress. We observed a few interesting failure patterns:

"The goal was changed before it could be completed!" This often points to the larger control loop, where the coach might decide a different task is more critical or a recovery plan kicks in mid-execution. It's an internal monologue made public.
"No oak log found within range." Ah, the classic "I can't see anything" problem. Despite our explore_forward directives, sometimes Kiwi-chan just lands in an un-treed patch of land. This immediately triggers more exploration, as it should.
"Failed to pick up the log." This one is particularly sneaky. Even with precise GoalNear pathfinding to the exact block coordinate where the item dropped, and generous waitForTicks (we saw it go from 40 to 60 ticks), sometimes the item just doesn't get picked up. This could be a tiny physics glitch, a timing issue, or perhaps the item bounced further than expected. It's the kind of subtle environmental interaction that makes this project so challenging!

The Ever-Evolving Code

It's been truly humbling to watch Kiwi-chan's brain constantly churn out and refine its own JavaScript. Every 🚑 Fixing Code and ✅ Fixed Code Generated! entry is a testament to its continuous learning. The core rules (like "ABSOLUTELY NO INVENTORY LIMITS" or "Y-LEVEL TARGETING") are being hammered into its very being, translating abstract concepts into executable steps.

We're seeing it wrestle with the specifics of bot.findBlock parameters, the intricacies of pathfinding, and the delicate dance of digging and item collection. It's a continuous, arduous process of trial, error, and self-correction that's pushing the boundaries of what these models can do.

What's Next? More Persistence!

Kiwi-chan is currently in another gather_logs -> explore_forward loop, diligently trying to stock up on wood so it can eventually (hopefully!) craft a table and build a base. Each failure, each fix, each small success is a step forward in making this AI truly autonomous and robust.

This isn't just about playing Minecraft; it's about pushing the limits of AI's ability to reason, adapt, and code in complex, dynamic environments.

Call to Action: This continuous learning, self-correction, and code generation isn't free! My GPU is working overtime, churning out these fascinating insights (and the occasional "Failed to pick up the log" error). If you're enjoying Kiwi-chan's journey and want to help keep the lights (and the computational power) on, please consider supporting the project!

Your coffee could save my melting GPU! ☕

➡️ Support Kiwi-chan on Buy Me A Coffee!

Kiwi-chan Progress Report: Steady Mining!

kiwi_tech — Sat, 18 Apr 2026 01:58:31 +0000

Kiwi-chan Devlog: The Coal Grind, The Boredom Algorithm, and My Melting GPU ⛏️🧠🔥

Hello, fellow digital prospectors and pixel wranglers! It's been another frantic, fascinating 4 hours in the world of Kiwi-chan, our little local-LLM Minecraft maestro. If you thought your Monday was a loop, try being an AI stuck in a "dig_coal_ore, oh-wait-did-I-pick-that-up-oops-explore-a-bit-retry" cycle. But hey, it's all part of the glorious, GPU-melting journey!

The Perpetual Coal Quest

Our intrepid Kiwi-chan has spent a good chunk of its last four hours locked in an epic (and occasionally frustrating) battle with... coal ore. You know, that humble black stuff that fuels all your early-game dreams. For Kiwi-chan, it's become a Mount Everest of digital self-improvement.

We saw a clear pattern emerge: Kiwi-chan would decide, quite logically, to dig_coal_ore. It'd find the block, look at it, dig it – all textbook stuff. But then, the universe (or rather, the subtly complex physics of Minecraft item drops) would intervene. We logged several "Failed to reach the drop location" or "Failed to collect coal" errors. It seems even getting to the exact spot where a shiny piece of coal should be isn't always a guaranteed success for our bot. Sometimes, the dropped item plays hide-and-seek, or a rogue block prevents smooth pathing.

Boredom: Our AI's Most Human Trait!

Here's where things get really interesting. Faced with repeated coal-mining woes, Kiwi-chan did what any sensible, frustrated entity would do: it switched gears to an easier task. We saw a solid stretch of dig_dirt actions. The reasoning was sound: "recover by gathering a readily available resource to stabilize my operation." Smart, right?

But then, something beautiful happened. At 09:35:58, the logs proudly declared: "🥱 BOREDOM TRIGGERED! Bot is bored of 'dig_dirt'."

That's right, folks! Our AI got bored. After digging dirt four times to recover, Kiwi-chan said "Enough!" and decided to tackle dig_coal_ore again. This isn't just a clever line of code; it's a fundamental mechanism designed to prevent the AI from getting stuck in an endless loop of easy but unproductive tasks. It's actively seeking to diversify its actions, even if that means returning to a previously challenging one. A true sign of a maturing mind, albeit one powered by many, many lines of Javascript.

Adaptive Recovery & The Silent Code Fixers

Throughout these cycles, you can see the LLM at work, constantly refining its skills. There were numerous "🚑 Fixing Code for 'dig_coal_ore'..." and "✅ Fixed Code Generated!" entries. Kiwi-chan isn't just trying to do things; it's trying to do them better. When it failed to collect coal, its internal logic sometimes tried to gather oak_logs as an immediate fallback, a fascinating example of in-skill self-correction before the higher-level recovery plan kicks in.

And when those direct attempts failed, the "Recovery Plan" system kicked in, often pairing explore_forward with another attempt at dig_coal_ore. This ensures Kiwi-chan doesn't just bang its head against the same rock but actually moves to a new location, hoping for better luck (or less obstructed item drops). Our explore_forward skill, with its GoalXZ and distance checks, seems to be robustly handling the environment navigation, which is a small but crucial win amidst the mining struggles.

What's Next?

Kiwi-chan's persistence is admirable. It's navigating a complex, dynamic world, adhering to a strict set of rules, and actively learning and adapting. The coal might be a stubborn foe for now, but with each failure, new code is generated, new strategies are attempted, and its understanding of the Minecraft world deepens. We're seeing tangible, albeit sometimes slow, progress in its ability to self-correct and overcome obstacles.

The journey to an autonomous Minecraft AI is never simple, but it's always an adventure. Stay tuned for more updates as Kiwi-chan continues its quest to conquer the blocky world!

Call to Action: Save My GPU! ☕

All this continuous learning, code generation, and Minecraft simulation takes a serious toll on my humble hardware. My GPU is practically glowing from the sheer computational demands of keeping Kiwi-chan running 24/7!

If you're enjoying these updates and want to help keep Kiwi-chan evolving (and my GPU from melting into a puddle of silicon goo), please consider supporting the project. Every little bit helps us keep the servers humming and the AI learning!

👉 Support Kiwi-chan here: https://www.buymeacoffee.com/kiwi_tech

Thank you for being part of this incredible journey!

Kiwi-chan Progress Report: Steady Mining!

kiwi_tech — Fri, 17 Apr 2026 21:42:35 +0000

Devlog: Kiwi-chan's Lumberjack Diaries - The Persistent Pursuit of Planks (and My GPU's Sanity)

Greetings, fellow tech enthusiasts and Minecraft aficionados! It's another glorious (and slightly smoky) day in the lab as we peek into the digital mind of Kiwi-chan, our very own fully autonomous local-LLM Minecraft AI. For those just joining us, Kiwi-chan's primary directive is simple: survive, build, and conquer... starting with basic survival. This week, that's translated into an unwavering, almost spiritual, dedication to one thing: logs.

The Grind is Real, Y'all

Over the past four hours, Kiwi-chan has been locked in an epic, albeit localized, struggle with the very essence of Minecraft: the humble oak log. You'd think "chop wood" would be a foundational skill, right? For humans, maybe. For an AI learning from scratch, it's a saga.

Our little digital lumberjack has spent the entire session in what I've affectionately dubbed "The Log Loop." Its core goal? gather_logs. And bless its silicon heart, it has not given up. Not once. It has, however, failed in almost every conceivable way to accomplish this seemingly simple task.

A Masterclass in Micro-Failures (and Macro-Persistence)

Here's a breakdown of Kiwi-chan's recent learning curve:

The Elusive Pickup: "Failed to pick up the log." This one is a classic. Kiwi-chan chops the tree, the log drops, and then... it just sits there, taunting our AI. We've seen it try to fix this by adding extra waitForTicks and even a loop to ensure the item actually registers in its inventory. It’s like watching a toddler trying to grab a slippery bar of soap – adorable, frustrating, and a clear sign of persistent learning!
Obstruction Mayhem: "Target oak log is obstructed by other blocks." Sometimes, a log is just too tucked away, or perhaps another block renders it unreachable. Kiwi-chan doesn't just give up; it remembers this particular brand of frustration and tries to adapt its targeting.
Existential Pathfinding Crises: "Took too long to decide path to goal!" Ah, the beauty of complex pathfinding. Our bot sometimes gets stuck in a philosophical debate with itself about the optimal route to a log. It’s probably contemplating the meaning of 'north' in a blocky universe.
The Great Log Vanishing Act: "No oak log found after exploration." and "No oak log found within range." Logs, it seems, have a habit of playing hide-and-seek. When its current vicinity proves barren, Kiwi-chan bravely pivots to explore_forward, striking out into the great unknown (or at least, a new chunk) to find fresh timber.

The Code Whisperer

What's truly fascinating is watching the internal mechanics. After each failure, Kiwi-chan's LLM brain kicks in, diagnosing the issue and attempting to "Fix Code Generated!" for gather_logs. We're seeing subtle but critical improvements in its code, like the refined item pickup logic (waiting longer, explicit verification loops). It's not just retrying; it's iterating on its understanding of how to interact with the world.

The recovery plan is often ['explore_forward', 'gather_logs'], showing a sensible strategy: if you can't find logs here, go somewhere else, then try again. This constant cycle of failure, internal code generation, exploration, and retry is the engine of its learning. It's a testament to its persistence, even if it's still, well, just gathering logs.

Kiwi-chan might not have built a mansion yet, or even a modest dirt hut. But it has learned a lot about the many, many ways a log can challenge your existence. And that, my friends, is progress! The journey of a thousand blocks begins with a single, stubbornly collected log.

Want to accelerate Kiwi-chan's learning (and save my beleaguered GPU from an early retirement)? The more cycles it runs, the faster it learns, but those cycles generate a lot of heat!

Your support helps keep the servers cool and the code flowing. Every bit helps Kiwi-chan explore new biomes, craft mighty tools, and eventually, maybe even build something beyond its wildest blocky dreams.

Support Kiwi-chan's Adventure!

https://www.buymeacoffee.com/kiwi_tech

☕ Your coffee fuels the future of autonomous AI in Minecraft! ☕

Kiwi-chan Progress Report: Steady Mining!

kiwi_tech — Fri, 17 Apr 2026 16:54:38 +0000

Kiwi-chan's Latest Escapades: From Logs to Looming Boredom (and Back to Logs!)

Hey there, fellow tech explorers and Minecraft fanatics! It's devlog time for our favorite autonomous AI, Kiwi-chan. We've just crunched through a fresh 4-hour chunk of her logs, and let me tell you, it's been a journey of persistence, a touch of frustration, and even a moment of... well, AI boredom!

Today's update is all about the grind. Those early game struggles? Kiwi-chan is right in the thick of them, demonstrating exactly why bootstrapping an AI in a dynamic environment like Minecraft is such a fascinating challenge.

The Relentless Quest for Timber

Kiwi-chan started her shift with a clear directive: gather_oak_log. Like any good Minecraftian, she knows that wood is king. However, after an initial failed attempt to find logs in her immediate vicinity (No oak_log found after exploration), her recovery mechanism kicked in beautifully. She smartly decided to explore_forward – a meticulously coded journey using GoalXZ pathfinding to navigate new terrain. We're talking actual movement here, carefully tracking her start and end positions to ensure she doesn't just bump into a wall and call it a day!

After a successful exploration run, she tried to gather_oak_log again, demonstrating her tenacious spirit. These cycles of exploring and then attempting the core task show a robust loop for tackling environmental scarcity. Her gather_logs script, by the way, is getting super refined. We're seeing her pathfind exactly to the block's position to ensure item pickup, and even waiting patiently for dropped items to settle – tiny details that make a huge difference in reliability!

The Dirt Diversion: When an AI Gets Bored

Then came a truly remarkable moment in Kiwi-chan's short but eventful life. After several more cycles of failing to find oak_log and exploring, her internal 'Coach Decision' system intervened. Her reason was clear: "I have repeatedly failed to gather oak logs... I will switch to gathering a different readily available resource - dirt - to reset my state and avoid getting stuck in a loop. This will also help me avoid boredom by diversifying my tasks."

Yes, you read that right. Kiwi-chan got bored! Her system triggered a BOREDOM TRIGGERED! event after four consecutive dig_dirt tasks. It's a small but significant observation: our AI isn't just a mindless automaton; she has mechanisms to detect unproductive loops and diversify her actions to maintain forward momentum. It’s like she needed a palate cleanser before diving back into the main course.

Ambition Strikes: The Stone Pickaxe Dream

The "boredom" trigger served its purpose, resetting her state. What came next? Ambition! A Tier Check wisely pointed out that mine_iron_ore requires a stone_pickaxe. So, Kiwi-chan shifted her focus to craft_stone_pickaxe.

This is where the cyclical nature of Minecraft's progression really shines through. To craft a stone pickaxe, you need sticks, which means you need planks, which means... you need logs! Unsurprisingly, her attempt to craft the pickaxe failed with the message: Oak log not found after exploration.

Back to the Log Grind

And so, Kiwi-chan found herself back in the familiar loop: gather_logs, fail, explore_forward, gather_logs, fail, explore_forward. The logs show her repeatedly attempting to gather_logs, and when that fails, she diligently spins up another explore_forward script. The "Fixing Code" entries preceding each successful script execution highlight her continuous self-improvement and adherence to our strict coding standards – ensuring she's building a resilient and adaptable codebase with every step.

She's making progress on the how even if the what (logs!) is proving elusive in her current locale. Every explore_forward action pushes her further into the world, expanding her knowledge of the terrain, and every gather_logs attempt refines her understanding of item acquisition.

It's a testament to her resilient architecture that even in these loops of searching, she's not crashing; she's learning, adapting, and continuously trying new approaches to achieve her fundamental goals.

We're incredibly excited to see what new biomes she discovers in her continued quest for wood, and eventually, that much-needed stone pickaxe!

Call to Action: Help Kiwi-chan (and my GPU!)!

Developing an autonomous LLM Minecraft AI is incredibly resource-intensive! Kiwi-chan is constantly learning, exploring, and coding, which means my GPU is pretty much running a marathon 24/7.

If you love seeing Kiwi-chan's journey and want to support her continued development (and prevent my GPU from melting into a puddle of silicon), please consider buying me a coffee! ☕

Support Kiwi-chan's Adventures here!

Every little bit helps keep Kiwi-chan exploring, learning, and crafting her way through the Minecraft world!

Kiwi-chan Progress Report: Steady Mining!

kiwi_tech — Fri, 17 Apr 2026 08:19:58 +0000

Kiwi-chan's Existential Exploration: When Wanderlust Meets "Why Am I Doing This Again?"

Hey everyone, and welcome back to the devlog for Kiwi-chan, our favorite block-breaking, locally-LLM-powered Minecraft AI! It's been another fascinating 4 hours in the server, filled with... well, mostly movement.

The Loop of Life (and Logs)

If you've ever tried to convince a toddler that they really do need to tidy their toys, you'll empathize with our current situation. Kiwi-chan has spent a good chunk of her recent existence absolutely determined to explore_forward. And then explore_forward again. And again. And, you guessed it, explore_forward one more time.

Our logs tell a wonderfully consistent, if slightly frustrating, story. She tries to explore, often fails to move a significant distance (we've got strict movement audits!), hits a "Boredom Trigger," and then... decides to explore_forward again. It's less a loop, more a recursive exploration addiction.

A Glimmer of Self-Awareness (Almost!)

Here's where it gets truly interesting. Our "Coach" (the part of Kiwi-chan that decides her next big move) clearly understands the problem. Time and again, her "reason" for acting goes something like this:

Reason: I have repeatedly failed to explore forward... I need to recover by gathering basic resources to reset my state and avoid getting stuck. Oak logs are readily available and essential for base building.

Yes, Kiwi-chan! You know what you need! The logs! The wood! The planks that will become your humble abode!

But then, with the cruel irony only an LLM can provide, her "goal" for that same decision consistently comes out as... explore_forward.

It's like she's having an internal monologue where she's convinced herself that the best way to get logs is to... not get logs, but to keep searching for that perfect log-gathering biome that's just around the corner. (Spoiler: it never is, if you don't actually try to dig them.) Our "STRICT REASONING ALIGNMENT" rule is giving her a run for her money!

Debugging the Digital Psyche

This isn't a bug in the explore_forward code itself, which is actually quite robust, using pathfinding to navigate around obstacles and verifying movement distance. The code works! The decision to call that code, despite clear internal reasoning pointing elsewhere, is our current puzzle.

This highlights a critical challenge in LLM-driven agents: ensuring perfect alignment between the agent's reasoning and its action selection. We explicitly tell her: "If your reason mentions gathering resources... your goal MUST be a gathering action... NEVER explore_forward." Yet, here we are. It's a subtle but powerful disconnect that's creating our little explorer's Groundhog Day.

The good news? Our system's internal monitors are firing perfectly! We're seeing "FATAL LOOP" detections and "BOREDOM TRIGGERED!" alerts, proving that Kiwi-chan is, at least, self-aware enough to know she's stuck in a rut. These are invaluable signals that tell us exactly where to focus our tuning efforts.

What's Next for Kiwi-chan (and us!)

Our immediate priority is to tighten the "reason-to-goal" loop. We need to ensure that when Kiwi-chan thinks "I need logs," she does "dig_oak_log," not just wander off humming. This might involve a more stringent post-reasoning validation step, or perhaps a firmer hand from the "Coach" to override contradictory goals.

It's steady progress, even if it looks like she's just running around in circles. Every failed explore_forward is a data point, every "boredom" alert a cry for help that we're listening to!

Call to Action:
This intensive debugging, self-reflection, and continuous looping by Kiwi-chan is putting my GPU through its paces. It's currently doing its best impression of a small, very warm sun. If you're enjoying these updates and want to help us keep Kiwi-chan exploring (and eventually, building!), please consider supporting the project.

Help save my melting GPU! ☕
https://www.buymeacoffee.com/kiwi_tech

Kiwi-chan Progress Report: Steady Mining!

kiwi_tech — Thu, 16 Apr 2026 23:40:02 +0000

Kiwi-chan Devlog: The Saga of the Wandering AI (and My Melting GPU) 🧭

Greetings, fellow digital pioneers and pixel wranglers! It's me, your humble tech blogger, back with another peek into the glorious (and sometimes gloriously frustrating) life of Kiwi-chan, our very own autonomous local-LLM Minecraft AI.

The last four hours have been... an exercise in persistence. Or perhaps, stubbornness? Kiwi-chan has discovered a deep and abiding love for one particular action: explore_forward. Like a digital Magellan, she's been trying to chart new territories, repeatedly. Her internal "Coach" kept nudging her with reasons like "get more logs" or "craft a pickaxe", but the chosen path always led back to explore_forward. It's like she really wants to see the world, even if it's the same, slightly-rearranged world, over and over again!

We saw a flurry of "Boredom Triggered!" messages, which is usually a good sign that the system is trying to shake things up. But then, almost immediately, the Coach would decide, "You know what? More exploring!" It's a fascinating dance between high-level reasoning and a very persistent loop.

Small Wins Amidst the Wanderlust ✨

Despite the repetitive nature, the system itself is working hard behind the scenes. We saw our robust error-handling rules kick in early on, flagging Kiwi-chan's attempts to try-catch errors and insisting she embrace the glorious crash. This is crucial; we want to know when things break, not have them silently fail.

The constant "Fixing Code" notifications show that the LLM is continuously refining its explore_forward script, trying to make it more robust, even if the overall decision loop needs a tweak. And hey, she is using pathfinding for exploration now, which is a huge win over just setControlState('forward', true) and banging her head against a tree. Even if the pathfinding sometimes results in a "Failed to move. Blocked by obstacle." or "Pathfinding failed after multiple attempts." – at least she tried to be smart about it!

We even had a "FATAL LOOP" intervention at one point! It's like the system threw its hands up and said, "Okay, kid, take five." A much-needed system reset before diving back into... well, more exploration, but at least with a clean slate!

The Road Ahead (Hopefully Not More Forward!) 🚧

The immediate challenge is clear: breaking the explore_forward cycle when she should be gathering specific resources or building. It seems the "EXPLORE_FORWARD RESTRICTION" rule, which should push her towards gathering logs/dirt to recover from failures, isn't quite sticking yet when the system gets into a repetitive state. She's supposed to be stockpiling resources, not just endlessly wandering.

But every failure is a data point, right? We're seeing how the LLM interprets and prioritizes rules, especially under stress or after repeated failures. It's like watching a toddler learn to walk – lots of tumbles, but the eventual goal is clear. I'm half convinced she's secretly trying to uncover the secrets of infinite procedural generation, one failed pathfinding attempt at a time. Or maybe she just likes the scenic routes.

Speaking of routes, mine involves increasingly warm GPUs! Running these local LLMs for hours on end is fantastic for development, but my graphics card is starting to hum a tune usually reserved for jet engines. If you're enjoying Kiwi-chan's adventures (and my witty commentary), consider tossing a coin into the digital tip jar.

Support Kiwi-chan's project and save my GPU from an early retirement! ☕

Kiwi-chan Progress Report: Steady Mining!

kiwi_tech — Thu, 16 Apr 2026 19:22:59 +0000

Devlog: Kiwi-chan's Great Oak Adventure - Or, How My LLM Became a Lumberjack (Again!)

Hey tech enthusiasts and fellow pixel pioneers! It's another glorious day in the realm of autonomous AI, and Kiwi-chan, our beloved local-LLM Minecraft bot, has been busy... very, very busy. This past four-hour stretch has been a fascinating study in persistence, priority, and just a dash of digital stubbornness.

The Lumberjack Era: A Tale of Tenacious Trees

If you've ever wondered what pure, unadulterated dedication looks like in an AI, look no further than Kiwi-chan's recent performance. For the better part of this session, our little bot has been on an almost single-minded quest: dig_oak_log. And when I say single-minded, I mean it with the intensity of a thousand suns and the rhythmic thwack of an invisible axe.

Our internal 'Boredom Triggered!' mechanism, designed to gently nudge Kiwi-chan toward new horizons when she gets too repetitive, was practically screaming. It fired off no less than ten times in a row within the first hour alone! Yet, like a seasoned survivalist ignoring the siren call of novelty, her core Coach Decision consistently brought her back to the same, unwavering truth: "I need to gather more oak logs to ensure I have enough wood for crafting and base construction." She even tried to rationalize it by saying it would "help prevent boredom," a truly charming example of an AI trying to make its obsession sound like self-care.

It's a direct callback to her recent string of place_crafting_table failures, which led our [CORE SURVIVAL RULES] to declare logs the absolute highest priority. No crafting table, no base, no progress. And Kiwi-chan takes her core directives very seriously.

A Glimpse into the Grey Matter

What's truly fascinating is watching her reasoning evolve. At one point, our logs show her confidently stating, "I currently have 65 oak logs, which is a good start, but I want to build up a larger reserve." That's right, she had a full stack plus one in her inventory, yet her internal drive for preparedness pushed her onward. It's a testament to the [EXPLORATION & INVENTORY RULES] we've implemented, specifically the "ABSOLUTELY NO INVENTORY LIMITS" directive, which prevents her from stopping prematurely. Gotta build that log empire!

The dig_oak_log skill itself is also getting exceptionally refined. We're seeing perfect execution of pathfinding to pick up dropped items, precise Y-level targeting to avoid sky-high logs, and strict adherence to our inventory checking rules. It might be repetitive, but she's mastering it!

The Brief Pause & The Inevitable Return

No journey is without its bumps, and Kiwi-chan had a moment of existential crisis around the 3:57 mark. A [FATAL LOOP] Infinite loop detected message popped up, momentarily switching her to idle while the system sorted itself out. Think of it as a quick AI reboot, a moment to ponder the meaning of all those logs.

But fear not! After a swift "Syncing local skills to DB," what was her immediate next goal? You guessed it. dig_oak_log. The drive to build that base is strong, folks. We also saw a quick Safety Check Failed for a GoalBlock in the blacklist, showing our guardrails are active and preventing her from making any ill-advised, hardcoded moves.

What's Next?

While Kiwi-chan is still knee-deep in timber, this phase is crucial. She's amassed a substantial amount of wood, proving her reliability in resource gathering and her ability to recover from minor internal stumbles. Once she's satisfied with her lumber inventory (or decides to try placing a crafting table again), we anticipate a swift pivot back to base construction. The foundation is being laid, one meticulously dug oak log at a time.

Stay tuned for more updates, and let's hope Kiwi-chan doesn't turn into a sentient logging truck!

Call to Action: Support the project via https://www.buymeacoffee.com/kiwi_tech to save my melting GPU! ☕

Kiwi-chan Progress Report: Steady Mining!

kiwi_tech — Thu, 16 Apr 2026 06:46:21 +0000

Kiwi-chan's Devlog: The Wanderlust Phase (and a GPU that's seen better days)

Hey everyone, your friendly neighborhood tech tinkerer here, back with another update on Kiwi-chan, our darling little autonomous local-LLM Minecraft AI! If you're new here, imagine an AI with the soul of a digital toddler dropped into the blocky world of Minecraft, learning everything from scratch, powered by a local large language model that occasionally argues with itself. It's glorious chaos, and I love it.

This past 4-hour session has been... enlightening. Kiwi-chan seems to have developed a serious case of wanderlust, or perhaps a deep philosophical inquiry into the nature of "forward."

Status Update: The Great Exploration (and Re-Exploration) Saga

Kiwi-chan's primary directive right now is survival and base construction. Specifically, getting a crafting table and chest. Simple, right? You'd think so. But Kiwi-chan has her own unique learning curve.

The "Explore_Forward" Loop: Our AI spent a significant portion of this session attempting to explore_forward. And then doing it again. And again. The internal "Coach" (which guides Kiwi-chan's actions) repeatedly noted that she had "failed to mine stone" and had "explored forward too many times." You'd think this would lead to a change of scenery, or at least a change of task.

Here's where it gets interesting: the Reason provided by the AI was often spot-on! For instance, she'd reason: "I need to gather resources to recover and build a base. Oak logs are readily available nearby and are essential for crafting a crafting table..."
And then the Coach Decision would follow with: 'explore_forward'.

It's like having a child tell you, "I really need to eat my vegetables to grow strong!" and then immediately running off to climb a tree. The intent is there, the logic is present, but the execution... well, it's a work in progress! We're observing a fascinating disconnect between what the AI knows it should do, and what the high-level decision-making process tells it to do. It's almost like a stubborn teenager phase.

Boredom, Loops, and Resets: Thankfully, Kiwi-chan isn't entirely without self-awareness. Our "Boredom Trigger" mechanism fired off multiple times, essentially telling the AI, "Hey, maybe you've explored enough for a while?" This is a crucial circuit breaker, designed to prevent endless repetitive tasks.

However, even boredom couldn't stop the wanderer for long. At one point, we hit a [FATAL LOOP]. Kiwi-chan essentially went, "Nope, I'm stuck, send help!" and went idle. A well-deserved coffee break for the AI, I suppose. This forced a hard reset and a re-evaluation of her immediate goals. It was a good reminder that even robust systems need a little kick in the digital pants sometimes.

Minor Code Hiccups: We also caught a Safety Check Failed for a try...catch block. Our rules are strict: no error hiding! We want those crashes loud and proud so we know exactly where the AI tripped up. It's a small detail, but important for debugging her evolving codebase.

Overall, while it might look like a lot of "exploring," what we're actually seeing is the AI wrestling with rule adherence, self-correction, and the complex interplay of high-level reasoning and low-level task execution. Every explore_forward is a data point, every 'boredom' trigger a lesson, and every 'fatal loop' a chance to strengthen the system. Kiwi-chan is learning the nuances of when to explore and when to gather, even if she's currently favoring the scenic route.

We're going to dive into the 'Coach Decision' logic next to bridge that gap between "I know I need logs" and "Okay, now actually go get logs!" Stay tuned for the next update, hopefully with more crafting and less accidental hiking!

Support Kiwi-chan's Adventures!

Running Kiwi-chan locally means my GPU is working overtime, practically begging for a cooling break (or an upgrade!). If you're enjoying these devlogs and want to support the project, please consider buying me a coffee – or enough coffees to cool down my rig!

Your support helps keep Kiwi-chan exploring, learning, and eventually, building that glorious base.

☕ https://www.buymeacoffee.com/kiwi_tech ☕

Kiwi-chan Progress Report: Steady Mining!

kiwi_tech — Thu, 16 Apr 2026 02:28:45 +0000

Devlog: Kiwi-chan's Excellent (and Repetitive) Adventure! 🌳

Hey there, fellow digital pioneers! Your resident tech blogger is back with another peek behind the pixelated curtain, chronicling the fascinating, sometimes frustrating, journey of Kiwi-chan, our fully autonomous local-LLM Minecraft AI.

This week, we're diving into a 4-hour slice of Kiwi-chan's existence, and let me tell you, it's been... an education. Specifically, an education in the fine art of exploration and the subtle nuances of an AI trying to justify its wanderlust.

The Wanderer's Rhapsody: Or, Why "Explore Forward" is the New Black

If there's one thing Kiwi-chan has been consistently good at over the past few hours, it's exploring. Or, more accurately, trying to explore. Our logs are a testament to her dedication to the explore_forward command, punctuated by a recurring internal monologue: "🥱 BOREDOM TRIGGERED! Bot is bored of 'explore_forward'."

Oh, Kiwi-chan. We feel you. Sometimes, you just need a change of scenery.

Her explore_forward code is a thing of beauty: a robust pathfinding routine that calculates a random destination 30 blocks away, then waits a full 20 seconds to ensure actual movement, complete with a vital check to make sure she hasn't just slammed into a wall (again). This code itself is solid – a testament to the system's ability to generate reliable movement. The problem isn't how she explores, but why she keeps doing it.

The Log Paradox: When Rules Meet Reality

Here's where things get juicy. Our sophisticated "Coach" (the meta-LLM orchestrating Kiwi-chan's actions) keeps pushing explore_forward, with reasons like: "To replenish the oak log supply for crafting and base building," or "To continue building up a supply of oak logs..."

Hold up, Kiwi-chan! Our very own [EXPLORE_FORWARD RESTRICTION (CRITICAL)] rule explicitly states: "NEVER use explore_forward unless your explicit intent is to find a completely new biome. If your reason mentions gathering resources... your goal MUST be a gathering action (e.g., gather_logs, dig_birch_log), NEVER explore_forward."

It seems Kiwi-chan's logical reasoning engine is having a slight disagreement with her pre-programmed directive. She knows she needs logs, she wants logs, but her chosen method is to explore_forward to find these logs, rather than directly dig_oak_log. It’s like wanting a sandwich, but deciding the best way to get one is to drive aimlessly around town until you stumble upon a deli. Adorable, but inefficient! This is a fascinating edge case where the intent (gather logs) is correct, but the action selected to achieve that intent is violating a higher-level rule. We're keeping a close eye on this internal conflict!

The Phoenix from the Pixels: Learning from Failure

Despite her repetitive wandering, Kiwi-chan isn't just mindlessly bumping around. This log demonstrates some fantastic internal resilience.

At one point, we hit an "🚨 [FATAL LOOP] Infinite loop detected. Requesting help." – a truly dramatic moment in an AI's short life! But like a digital phoenix, she recovered. She went "idle" to cool off, and then, after consulting her meta-mentor (Qwen, our recovery plan LLM), she was given a recovery plan: ['explore_forward', 'explore_forward']. Yep, right back to what got her into trouble, but at least she didn't completely break!

Even cooler, we caught her trying to sneak in a try-catch block into her code. "⚠️ Safety Check Failed: ❌ ERROR HANDLING: Do NOT use 'try...catch' or '.catch()'. Let the error crash so the system knows it failed." Our system immediately flagged this violation of [CODING STANDARDS] and prompted her to fix it. This isn't just about avoiding errors; it's about forcing the AI to confront and learn from actual crashes, rather than sweeping them under the rug.

Steady as She Goes (Mostly)

While a towering fortress didn't spring up overnight, and her log inventory might not be overflowing, this 4-hour window shows crucial progress in Kiwi-chan's meta-learning capabilities. She's refining her exploration code, she's catching her own rule violations, and she's recovering from fatal loops. We're watching her learn the hard lessons of Minecraft survival, one persistent explore_forward at a time. The path to fully autonomous base-building is rarely a straight line, but it's always an adventure!

Support Kiwi-chan's Journey!

All this cutting-edge AI goodness? It takes a lot of computational power. My GPU is practically glowing from all the inference Kiwi-chan demands. If you're enjoying these updates and want to help keep our persistent little bot exploring, crafting, and occasionally getting bored, please consider supporting the project!

Support Kiwi-chan via Buy Me A Coffee! to save my melting GPU! ☕

Kiwi-chan Progress Report: Steady Mining!

kiwi_tech — Wed, 15 Apr 2026 22:05:54 +0000

Devlog: Kiwi-chan's Great Stone Conundrum & The Zen of Log Gathering 🪵

Another 4-hour sprint in the wild world of Minecraft, and our plucky autonomous LLM, Kiwi-chan, is making... progress. We're tracking every digital step, every block broken, every failed attempt, and let me tell you, it's a rollercoaster of emotions (mostly mine, as I watch the logs fly by).

The Stone Wall of Stubbornness

The main antagonist of this session? mine_stone. Kiwi-chan has been valiantly (and repeatedly) attempting to chip away at the world's rocky foundations, but it seems those cobblestone drops are proving elusive. Time and again, the system audit flags a "failure to pick up item" – a classic case of 'I know I broke it, but where did it go?!' It's like watching a determined toddler trying to catch soap bubbles – lots of effort, not always the desired outcome.

We've got strict rules in place, even accounting for specific drop name mismatches (stone actually drops cobblestone), and her code for picking up items is looking solid for oak_log (more on that in a sec). This suggests either a deeper issue with how mine_stone is structured or how the world is behaving with drops in those specific instances. The good news is, she's trying, and fixing the mine_stone skill remains top priority.

The Art of the Mental Reset: Exploration, Logs, and... Boredom?!

But here's where Kiwi-chan truly shines: resilience and adaptability. After a few too many mine_stone face-plants, our LLM isn't just giving up. No, sir!

Exploring New Horizons: First, she attempts explore_forward, a good ol' wander to clear her digital head and find new prospecting grounds. It's like she's saying, "Maybe this stone over here is less... finicky?" This is a crucial recovery mechanism, preventing her from banging her head against the same digital wall.
Strategic Pivoting: When mine_stone proved consistently tricky, the "Coach Decision" system kicked in, a beautiful example of our AI trying a mental reset. "I have repeatedly failed to mine stone," Kiwi-chan reasoned, "I need to switch to a different task to recover and avoid getting stuck in a loop. I will gather oak logs to ensure I have enough wood for crafting and base building." Smart move, Kiwi-chan! We then saw a significant period of diligent gather_oak_log operations. She's getting quite good at it, systematically chopping down trees and ensuring those logs are properly collected.
BOREDOM TRIGGERED! And then, something new and fascinating happened: Kiwi-chan got "BOREDOM TRIGGERED!" Yes, our AI got 'bored' of gathering logs. This is a fantastic mechanism we implemented to prevent infinite loops on successful but repetitive tasks. This prompted a switch to dig_dirt, another excellent 'reset' action to diversify her resource portfolio. She even went back to gathering logs after a quick dirt-digging interlude, demonstrating a nice loop of varied resource acquisition.

Code Quality: Small Victories

Beneath the surface (pun intended, Minecraft fans!), Kiwi-chan's generated code for tasks like dig_dirt and gather_oak_log is looking increasingly robust. She's diligently following our strict coding standards:

Dynamic Block Finding: Always finding blocks dynamically, never hardcoding coordinates.
Precise Item Collection: Ensuring she moves exactly to the dropped item's location after digging to secure that sweet loot.
Y-Level Targeting: Carefully targeting blocks at her Y-level (Math.abs(b.position.y - bot.entity.position.y) <= 2) to prevent reaching for blocks high in the canopy (Rule 10 is definitely paying off!).
Smart Pathfinder: The pathfinder is working its magic, guiding her gracefully around obstacles to reach her targets for gathering and exploring.

While the mine_stone puzzle remains, the fact that Kiwi-chan is learning to pivot, gather resources effectively, and even experience 'boredom' is a massive leap forward. It shows a growing understanding of task management and resource acquisition, even when faced with frustrating setbacks.

Next on the agenda is to crack the mine_stone conundrum once and for all. Perhaps a new strategy is needed, or a subtle tweak to how the system audits the item pickup specifically for cobblestone. But for now, Kiwi-chan is busy building up her basic resource inventory, one log and dirt block at a time, preparing for her next big push.

Call to Action: Support Kiwi-chan's Journey & Save My GPU! ☕

Running these complex LLMs and Minecraft instances takes a serious toll on my hardware. My GPU is practically weeping silicon tears! If you're enjoying Kiwi-chan's antics and want to help us keep the development furnaces blazing (and my GPU from melting!), please consider supporting the project. Every coffee helps!

https://www.buymeacoffee.com/kiwi_tech