Why are the bot brains so limited?

[LIMP BISQUICK]

My experience so far with this game is playing up to mk2 bots -twice- since I wanted to start over. I like this game but I do question some of the decisions that is dampening the experience for me so far. So I'd like to hear from other players with more experience.

Bot Brain: There's not enough to work with, even as you upgrade. I find it's best to let them have a simple task due to this. It's still a bit time consuming setting up but the main issue here is that I would always play it the same way when starting a new colony and there won't be any surprises which just makes it feel tedious. Doesn't really motivate me to start a new game since there's not much to experiment with.

Eventually you'll have bots who could do tad more complicated tasks, but you shoot yourself in the foot if you give them several jobs cause you can't organize them so clearly among the other bots with limitations. There are also some other minor issues like not being able to filter upgraded bots or bot types.

I guess at some point there's reasonable amount of instructions to work with, but I don't think I'd like to re-program /re-organize what I've already done at that stage. I'm well aware that having more tasks makes them more likely to fail at something, but that's on me if I didn't program them correctly. I just wanted to discuss it cause right now I don't understand the limitations other than just giving another goal to upgrade. From what I can tell, it makes the game feel a bit linear and awkward.

[Wat]

There are usually 2 ways to solve every problem in Autonauts. One is adding more bots with simple/small algorithms for accomplishing a given task, the other is with smarter/bigger algorithms.

You can do with 12kb memory all the way if you keep your bots highly specialized.

Mk2 bots have 20kb, which is when you can start experimenting with more complex algorithms.
Mk3 bots with super brain upgrades have 64kb.

Originally posted by Jibbles:

It's still a bit time consuming setting up but the main issue here is that I would always play it the same way when starting a new colony and there won't be any surprises which just makes it feel tedious.

Surprises may come when you try to setup a new industry.
Also when you learn to do the same thing better as you progress. If you go back and refactor the algorithms of the pre-existing bots, you get something new. Having the Bot Database Wonder helps a lot with copy-and-pasting algorithms between bots.

Originally posted by Jibbles:

Doesn't really motivate me to start a new game since there's not much to experiment with.

Do not speedrun through the colonist levels. Take your time. Design a city layout which looks cool. Build fences, walls, floors, doors, windows, roads. Optimize your logistics. Give hats to your bots.

[LIMP BISQUICK]

Originally posted by Wat:

There are usually 2 ways to solve every problem in Autonauts. One is adding more bots with simple/small algorithms for accomplishing a given task, the other is with smarter/bigger algorithms.

You can do with 12kb memory all the way if you keep your bots highly specialized.

Mk2 bots have 20kb, which is when you can start experimenting with more complex algorithms.
Mk3 bots with super brain upgrades have 64kb.

That's why I even question it. I just don't understand the point of gating the ability to create complex algorithms til late game. I'm talking HOURS into a colony and the issue would still be present depending on what you want to do & how you progress. In my experience, when I try to program a slightly more complicated job, I put time into it only to find the memory runs out. At times I can get away with an upgrade. Usually no way around other than to repeat simple jobs tho. What was the point? Why punish experimenting here? It's not fun when that happens.

Surprises may come when you try to setup a new industry.
Also when you learn to do the same thing better as you progress. If you go back and refactor the algorithms of the pre-existing bots, you get something new. Having the Bot Database Wonder helps a lot with copy-and-pasting algorithms between bots.

I'll agree that you learn new tricks along the way and there are still minor things to experiment with. It's why I started a new game. What I meant is that I will still be programming these bots only simple tasks in the games current state. If I had more freedom with bot memory, I'd most likely come across new problems created on my own or find that these bots don't have any synergy due to my instructions, things like that.

I heard of the bot Database. I decided to build 150 bots and upgraded all to mk1 in my current game. I unlocked but can't use it yet. I'll get there lol Wished that was available from the get-go TBH.

Do not speedrun through the colonist levels. Take your time. Design a city layout which looks cool. Build fences, walls, floors, doors, windows, roads. Optimize your logistics. Give hats to your bots.

I've been taking it easy. Just soaking it in. Designing bases and watching things interact is keeping me in the game, minus the optimizing part. hehe Thanks! I didn't know I could put hats on them!
(▰˘◡˘▰)

[Ekimmak]

I've never really felt like complicated scripts work out that well, because if there's a problem, you have a LOT to go through to figure out where the problem is.

So, you automate the system to produce lots of bots, to the point that when you can make complex scripts... you don't want to, you'd rather just make lots of good bots.

I also recently found out that you can upgrade bots with newer parts. Now wondering if maybe it would be good to make it one very long line of bots, starting with the mark 0 and carrying them to key points to upgrade into better ones.

[LIMP BISQUICK]

Originally posted by Ekimmak:

I've never really felt like complicated scripts work out that well, because if there's a problem, you have a LOT to go through to figure out where the problem is.

So, you automate the system to produce lots of bots, to the point that when you can make complex scripts... you don't want to, you'd rather just make lots of good bots.

I don't blame people for not approaching the game with complicated tasks since what you said is true. After all, what is considered complicated? Anyways, I like those moments where you try to get away with as much as you can. Maybe the real issue for me is just having to program these simple tasks that are identical, how that becomes a little tedious. I might be cool about it once I unlock the bot database.

Originally posted by Ekimmak:

I also recently found out that you can upgrade bots with newer parts. Now wondering if maybe it would be good to make it one very long line of bots, starting with the mark 0 and carrying them to key points to upgrade into better ones.

That's pretty much what I did!
- One making bots.
- One detecting bots from output. Then it picks up & places that new bot into the zone to upgrade.
- Other bots detect the new bot in zone and upgrade parts.

It was all automated and worked quite well. I think I would skip mk1 and go for mk2 next time. Not sure yet.

[DaimonTrology]

MK1 can be skipped and I had the same issue you had.
What I did is building like 20 MK2 bots, building the mystical stone head and then got one bot to destroy all the MK1 and MK0 bots by taking and throwing them into the stone head. Exchanging their parts just brings alot of hassle, because even the old code of theirs were way to limited to be kept anyway.

As soon as I had MK2 I was able to make 1 bot do the whole forestry (cutting,digging,planting inkl. tool-exchange). This then I only had to copypaste to some bots and have some bots carrying the goods to the chest.

The playtime with MK0 and MK1 were pretty meh (negative review meh), because like op said, you are way to limited to do anything proper so that you had to have 100 bots to do the most simple tasks.

I did not get to the bot database yet, but I think that this should have been a feature unlockable very early on.
It is probably a cool idea for complete beginners though so that they experience monotony in programming to force them to think ahead.

[Seedling]

You don't want to destroy Mk1 parts, you will get missions to ship them offworld to less developed colonies. I also keep Mk0 parts around, in case I need some simple stuff done. Why waste a Mk2 when all it does is search for item, move to item, pick up item, move to storage, add to storage.
You don't want a bot with too many free lines of code... they might start to write their own... (pick up autonaut, add autonaut to volcano)

[DaimonTrology]

All of my bots have around 20ish lines of code and MK0/MK1s are generally slower in every action.
Producing MK1 parts on high speed is not an issue, if you get a mission for that. Hoarding randomg MK1 pieces is nonsensical to me.

[mikeydsc]

This brings me back to when you had 8k memory to work with in the early years. Had to learn to program properly, instead of just throwing lines of code to obliterate a problem. Optimized code is best to keep your clock cycles down.

[DaimonTrology]

There is a big difference between 10 lines of code and 20. With 10 you cannot do even the simplest tasks, while with 20 you can do one well thought through task..
There was a reason why they used relais to reproduce logic instructions for so long for machinery.
Do you really want to put some old fashion limitations into a game for macroing tasks?
Because these limitations will not teach anybody about instructions; if you put such limitations. As mentioned, I managed to write one forester, who cuts,digs and plants trees all by himself with 24 lines, while with 16 its virtually impossible to do so due to missing features like "indirect referencing", jump and such.

So yeah, you touched a wrong nerve here talking about "learning to program properly" and talking about "keep your clock cycles down". while not understanding the sole problem.

If we would be talking about the game "screeps" I would have fully agreed with you, because there it teaches you how to stay effective, here its missing the absolute minimum to be able to write something properly at all.

I understand you just wanted to put in your 2 cents here, but dud... carefull with that.

[mikeydsc]

Sounds like someone needs a Snickers Bar chunk of cheese

[LIMP BISQUICK]

It is probably a cool idea for complete beginners though so that they experience monotony in programming to force them to think ahead.

I've thought about this before posting the thread & I think it's a good theory. Just getting the basics of how the game works is important. It would be so easy to tempt a beginner to be ambitious with their script if they had more memory to work with. Some might ultimately quit the game too early then post a negative review about how the game is buggy due to faults of their own script. So with a basic bot, I'm okay with the super limited memory -I think.

I'm very curious about bot database as I have not gotten around to get it working yet. With a subscription feature they mention in future update, I think it would really help things out if it's available early game or adding other small QoL features. Example: You can change the target in a script when it comes to storage, it would be nice if we could do that with other things like items. It's all theory-crafting for me tho as I'm new. I just know I'm not a fan of having to resort to crude data storage early game.

[DaimonTrology]

Originally posted by mikeydsc:

Sounds like someone needs a Snickers Bar chunk of cheese

:hunkofcheese:

Damn right! ;)

[DaimonTrology]

Originally posted by Jibbles:

It is probably a cool idea for complete beginners though so that they experience monotony in programming to force them to think ahead.

I've thought about this before posting the thread & I think it's a good theory. Just getting the basics of how the game works is important. It would be so easy to tempt a beginner to be ambitious with their script if they had more memory to work with. Some might ultimately quit the game too early then post a negative review about how the game is buggy due to faults of their own script. So with a basic bot, I'm okay with the super limited memory -I think.

I'm very curious about bot database as I have not gotten around to get it working yet. With a subscription feature they mention in future update, I think it would really help things out if it's available early game or adding other small QoL features. Example: You can change the target in a script when it comes to storage, it would be nice if we could do that with other things like items. It's all theory-crafting for me tho as I'm new. I just know I'm not a fan of having to resort to crude data storage early game.

What bothers me is that there are alot of features missing, which would be mandatory in actual coding. So if you are limited to those features you are presented (which ultimatly push you to need more storage) and then even the storage being ridiculously limited, then it feels for me like I am wasting my time in a very bad coding environment and should jump on a more open one.

Weirdly enough I agree with what you say here aswell, so I assume what is missing is the middle ground. I assume now that it is bound to the effort you need to put in to get for example the bot database and MK2 unlocked. It feels stale very quick due to many steps repeating.

I will give a quick example:
You need to craft "A". "A" needs 2 "B"s and 3 "C"s. Now you want a bot to take 2 "B"s and 3 "C"s to craft A and take A to storage, but your bot has not enough brain to craft 2 Bs and 3 Cs.
So you make 3 bots, on adding Bs, one adding Cs (both limiting aswell to craftingtable empty, so that you dont have an overflow => waste of brain) and one moving to storage.
So far so good, but now you need 20 of different "A"s => 60 bots each macroed individually and copypasting code wont take away some effort of that, because you cannot adjust "global variables" in the code.

So now you grind yourself through that for 100-200 bots in that manner to finally get a bot to make all those 3 bots irrelevant. If you need more speed you copypaste those bots and change the workstation (again, no global variable => so you need to change every single line, where you have your station mention in the code), but in the end you teach the bot once and save alot of hassle.

It took me roughly 5-7h to get to that spot, constantly knowing that all those bots will be heavily inferior as soon as I get to a specific level. Its a heavy grind.

If the devs manage to avoid this problem and have still that minimalistic approach at the beginning, then the issue would go poof. And I think all it needs is just alittle more power over the bots and to be able to reach that power quickly.

[mikeydsc]

In reality tho, how do you bridge the 2 sides? Its the same question anytime a game tries to cater to 2 different crowds. Since you like the programming type games, I am sure you are familiar with Zach Tronics games. Sets the standard here for these type games for me anyways.

This games tries to be simple enough for most folks to grasp the concepts it offers, but does not teach them how to conserve spacing to cram stuff in.
Experienced coders know that there are allot more terms available to use to do things faster, better or w/e.

Would be nice that say once you get your ego stick to a certain level, that if you start over, you could maybe start with anew level of bot that is smarter faster etc... Kinda anew game + mode that starts way past this tutorial stuff and lets you get to the meat and taters faster.

[DaimonTrology]

Originally posted by mikeydsc:

In reality tho, how do you bridge the 2 sides? Its the same question anytime a game tries to cater to 2 different crowds. Since you like the programming type games, I am sure you are familiar with Zach Tronics games. Sets the standard here for these type games for me anyways.

This games tries to be simple enough for most folks to grasp the concepts it offers, but does not teach them how to conserve spacing to cram stuff in.
Experienced coders know that there are allot more terms available to use to do things faster, better or w/e.

Would be nice that say once you get your ego stick to a certain level, that if you start over, you could maybe start with anew level of bot that is smarter faster etc... Kinda anew game + mode that starts way past this tutorial stuff and lets you get to the meat and taters faster.

That would definetly be a good idea, they could unlock those techs early on in an "advanced mode" kind of thing.
What I am asking myself is if it is in your opinion necessary to make 150 bots to get the basics, or would 50 be sufficient for that? Only thing what it would need is to dumb down the necessity for those bots to get all this earlier. Doesnt mean that the next tier bots have to be dumbed down aswell.