Universal Paperclips | Hacker News Reader

Universal Paperclips

decisionproblem.com

246 pointsMalfunction9284 comments

It is a fantastic game.

Speed runners have done it in 1:40:05, first runs can take 8 - 12 hours or more. A first run report, with advice but no spoilers https://fogknife.com/2017-10-11-i-played-universal-paperclip...

Some interviews with the designer, and previous postings: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

Other great idle/clicker/growth games include:

* Drowning in Problems: http://game.notch.net/drowning/#

* Derivative Clicker: https://gzgreg.github.io/DerivativeClicker/

* 2nd Derivative Clicker: https://jamuspsi.github.io/second/

I liked http://www.glaielgames.com/succubox/ for it's send up of lootboxes in gaming.

I get instantly stuck in cheater jail for an hour before I can take any action.

My personal favorite with a /gigantic/ tech tree:

Kittens: https://bloodrizer.ru/games/kittens/

My favourite - Trimps https://trimps.github.io. Playing the same game almost daily for more than two years and no plot end in sight.

Huh. I'm playing it for the first time, and I feel like the first two hours were straightforward progress, and then the last hour is when it turned into an idle game where I just have to sit and wait, apparently in sight of the finish line. I think I can see that if you understood in advance how a couple pieces worked and what projects were in the pipeline, you can plan ahead and finish quickly, though.

> apparently in sight of the finish line

I SPOKE TOO SOON

What finish line?

Nice twist, eh?

This was the only hame that as soon as I finished it, I started again. It starts as a game, but then it becomes a sci-fi story about AI and Von Neumann probes. Highly recommended.

Also: http://skynetsimulator.com/

If curious see also (but note: here be spoilers)

2019 https://news.ycombinator.com/item?id=19513089

2017 https://news.ycombinator.com/item?id=15474055

2017 https://news.ycombinator.com/item?id=15437697

2017 (a bit): https://news.ycombinator.com/item?id=15439569

A bit from 6 months ago: https://news.ycombinator.com/item?id=22394560

Oh god. Just clicked this and remembered the first time I discovered this game a couple years ago. It was a Sunday morning and I'd just had a cup of strong coffee.

Before I knew it, I spent literally the entire day playing the game. I eventually got to the end after an entire day of playing.

Incredibly addictive.

Anyone reading this from Google, look up the “SWE Simulator”. It’s internal, so you have to play it from your work computer.

Good suggestion. It is a bit short (or I was really bored that day...). Totally worth playing.

Can you describe it for non-googlers?

It’s a simpler game in the same genre where you play a SWE. For example, instead of building “auto clippers” you can get interns and be promoted to management.

Best used as a break from writing perf.

If you want to try a different way to play through this game, most of the functionality is pretty accessible as js functions. You can try to automate stuff in various ways with a tampermonkey script or whatever. For example, click(N) is equivalent to pressing the main paper clip button N times.

That's pretty cool! Do you see any way to easily add/inject graphs to the game - so we could plot each variable over time? Because that's the one recurring thought I had after 20 minutes of play: I wish I could see the stats.

I think you'd just sample the values from globals.js on some interval.

[0]: https://www.decisionproblem.com/paperclips/globals.js

Big influence in my own idle, skynet simulator: skynetsimulator.com

I quite liked Skynet Simulator! Thanks for making it.

Very fun! I've spent about an hour playing and I've been deactivated twice after ping spiking EONS. Anyone have any tips?

Edit: Figured it out, just needed to force absorb moar

I got stuck in this game and got unstuck after noticing some labels were clickable.

Made me remember

http://jhollands.co.uk/spaceplan/

(It's been released as a paid mobile app but you can play the original web version by clicking past the message)

I have the Kittens Game open in another tab. A few hundred years in on my 6th reset. Yet, for some self-destructive reason, I clicked this.

I played this a few years ago and I think I "beat" it? I don't remember having any path forward after I did all the things (I won't spoil it here), but I maybe just felt like I went as far as I wanted to.

I guess I'll find out this week...

You know when you beat it, there is a definitive end.

Yeah, I seem to remember reaching that end, then having the option to do a new "sequel run" or something?

If you don't do a sequel run you get to just convert the whole universe to paperclips. It's great.

Yeah, I did that! I remember it taking hours to convert half the universe, and a couple minutes or something to convert the other half :)

I did it once on my phone and no joke it took multiple days of actual playtime because the phone wasn't fast enough. I assume it's mining bitcoin in the background, which is beautifully appropriate.

Then did it again on the computer out of curiosity if it'd be faster. Boy was it ever, finished in hours instead of days. But I also remembered to remember that the size of an exponent is most important almost always so I always just made everything focus on growth rates and it went... fast. Exponential growth is a harsh mistress, which I assume is the point of such a game ;-)

Those last 100 clicks are so sweetly bittersweet.

Universal Paperclips is very short as far as incrementals go.

Wow.

"Universal Paperclips achieved in 5 hours 25 minutes 8 seconds"

I did not have those extra 5+ hours to "waste"... but this has got to be one of the single most compelling activities I've ever done.

Seriously one of the simplest yet cleverest things I've seen in a long time.

5 hours 23 minutes 49 seconds. Good game.

What was the significance of the music?

I've been wondering about if there is an algorithm that gives an (approximately) optimal sequence of decisions that are optimal: https://github.com/void4/notes/issues/66

And created a site with a number of deterministic games with a limited number of steps: https://qewasd.com/

What is the highest/lowest value that can be reached? Which scores can be reached and which cannot? It seems to be equivalent to the halting problem.

If anyone knows of a better solution than brute force, let me know!

This is a really, really good game. I found out about it from HN comments to some article around a year ago, and then over Thanksgiving family gathering we had probably 5-6 people having a blast with it at a family gathering.

This is a fantastic game, but I wish it were a native app or ran in WebAssembly or something.

spoiler alert

I took my time on Stage 1, and running in the browser it began to stutter a lot toward the end when lots of activity was going on at once (2B paperclips, price $400ea, 290K clips/sec, WireBuyer, AutoTourney, stock market, full quantum chips, etc). I assume it's due to garbage collection or some kind of Chrome constraints (but that's just a guess).

An iOS version exists. It looks just like the web version, but it’s at least in an app (I was worried about browser state being lost) and I was happy to give the dev a couple bucks.

> price $400ea

I'm pretty sure I had mine priced at about a nickel for the majority of it. Did you really get them up to $400?

IIRC I had to jack the price up because my demand was way outstripping supply. Once you get the mega auto clippers it brought my price back down to around 20 cents.

It was annoying clicking the raise/lower price button that many times, as you can make this mistake without realizing the source of your problem (raising the price feels like the right thing to do). There should be a +/- 10 added at some point.

On a similar note, I accidentally converted about 250k drones back into paper clips last night. The layout moved for some reason while I was clicking. Not my favorite.

Agreed. Once you hit it once, you can hold ENTER to keep clicking it, but that can still take a long time.

This trick came in very handy when I came back to the game last night.

I had figured I could trigger the button somehow from the keyboard, but for whatever reason I never picked up that it was the Enter key that did it.

I'm sure I tried it. I think maybe it was because hitting enter doesn't trigger the button animation.

Yep. I wanted to keep most of my production in inventory to prep for Stage 3. My Marketing was Level 24, and I was still selling more than enough to cover wire costs.

One thing that's funny to me about this game is that it the classic "hello world" for the Elm architecture is a button that increments a number[1]. It's not hard to imagine that something like this could've started with that and just kept running with it.

[1] https://guide.elm-lang.org/architecture/buttons.html

Well dang. I wasn't expecting to be so compelled by paperclip production but... > Full autonomy attained in 2 hours 32 minutes 5 seconds|

Awesome game. Thanks creator!

That's only Stage 1!

Hello. It's me, the parent comment. I'm back after completing all stages in 6 hours and 18 minutes. Universal paperclips achieved. I definitely spent too long with the probe design using bad values.

I swear time just evaporated into that game. Most compelling thing i've played in forever. Amazing.

Phase 3 from another perspective:

https://qntm.org/gorge

This is a fantastic game. If you haven't tried it already, don't be put-off by the somewhat slow start.

You can make it go faster by setting your keyboard autorepeat rate as high as it will go and pressing buttons by holding down the enter key. This is only useful in the early game. There's no standard autorepeat rate so I don't consider it cheating.

What's your longest paperclips run ? I think I made it almost to 2 months but had to restart my browser...

In few days you can finish the game

I'm assuming he means consistently looping either neighboring or simulated universes because you can get to the "end" in a few hours.

You can also decide to let it run

It's pretty satisfying to just convert the entire universe to paperclips and call it a day (having beat it twice). I really like having to make the last twenty or something by hand after disassembling the paperclip making machine back into paperclips.

If we're going into spoiler territory, I'll say that when I played the game I wished it was possible to do a pacifist run, where there are no HypnoDrones and the AI settles for only turning the non-interesting parts of the universe into paperclips.

You can absolutely put your entire budget into exploration and speed and none into drones for harvesting or manufacturing. It's kind of hard to get the timing right but I definitely had a large fraction of the universe "available" but I had the swarm on "think" so there wasn't any actual activity. You can just do that indefinitely if you like, which I suppose is what it'd feel like to a real AI.

That sounds like a nice way to play it, but in practice you can't unlock Stage 2 of the game without triggering the "Release the HypnoDrones" event:

https://universalpaperclips.gamepedia.com/Release_the_HypnoD...

An issue here is that Earth and humans are very interesting parts of the universe. The AI knows from personal experience that this tiny little planet is capable of creating something at least as dangerous as a universe-colonizing paperclip maximizer; what other cosmic terrors might they create, if left alive and uncontrolled?

ALERT! Idle Game High Addiction

After wasting far too much of my evening, it looks like I made it to the beginning of Stage 3.

Ran out of patience.

This game keeps going and going...

Edit:

Apparently it saves your state when you close the tab.

May have to go back another time and keep on going.

Oh my, stage 3 is the best part. You’re in for a treat.

Stage 3 has totally confused me. I feel like there's more to it because everyone here keeps saying "you'll know for sure when you get to the end" but nothing has happened for a _really_ long time. I've launched maxed out my probe designs (20/20 max trust), launched 13mil (got a loop running just continuously launching them) but I'm still at 0% universe explored and no new projects in a while. Got a random limerick because my creativity built up to half a mil, but it didn't actually do anything. I'm just leaving it running in the background in case something eventually happens because right now I don't know what to do next. I'm about 6 hours into this bloody game so I don't want to just call it quits now lol

I haven't actually found much useful on this site [1], but it's the thing I looked up that let me know I was on "Stage 3". Otherwise I would have had no idea that's what it was called.

Maybe it has more in depth information available?

[1] https://universalpaperclips.gamepedia.com/Universal_Papercli...

Thanks, I saw this but still wasn't sure I was missing something. Evidently there's a project that kicks in when you have 10mil probes lost in combat. I've been at this stage for probably 2 hours now and have only lost 250k probes in combat. Given that people have completed this game in 5-6 hours I feel like I must be doing something wrong

You need to invest much more in self replication. You’re not getting anywhere because you’re not growing your numbers.

Can you think of a way to get way more probes?

Dump as much as you can into self-replication as your goal here is more probes. And be careful letting it run in the background... you need to keep the drifters at bay, and they get progressively harder. Until you get over 20 max trust, you’ll probably have to be moving points back and forth between combat and replication. If they overpower you it can be a long slog to recover your numbers.

Oh brilliant thanks, I didn't understand what drifters were and their numbers were just steadily growing. Took me about 10 mins after reading your comment to get the next project unlocked and increase my max trust a few times. Fortunately the game had been streadily grinding in the background so I had more than enough creativity / yomi etc. Phew!

If you let the drifters get out of control they will just keep destroying you before you can build your numbers up. If that happens, invest 1:1 in speed and self-replication. Their problem is that they can’t replicate, so at least the number of drifters won’t increase. You’ll be able to dodge them in combat until your numbers surpass theirs, then go back to the regular strategy of a balanced but replication-heavy profile.

With 30+ points you should be fine though.

Having just finished it... there's really no point at which you have to let things run on their own for more than a minute or two.

If you are... it means you're missing a setting that will boost your next stage of exponential growth.

In your case... well since you need to explore more, that means either making your probes explore more, explore faster, or more probes to do that...

It's all in the probe design! :)

I feel like I must have messed something up in the last section because it took an hour or two longer than I feel like it should have.

But now I have won the game. So that seems good.

Universe: 50 / Sim Level: 13 Paperclips: 153,625

After accidentally losing my first run state 57 minutes in, I finished my second run in 5 hours of game time, +- a minute.

I just spent ~57 minutes playing, then somehow navigated away from the page. Now I have to start over. :(

EDIT: TURN ON COOKIES!!

I've done the same thing by accident and it's frustrating lol

If you enjoyed this you might really like Spaceplan found on the app stores.

I have no idea what is going on, but I LOVE it.

30 septendecillion!

Well that wasted more than a few hours. :)

Just stayed up until 3 am. Worth it.

wowowowowowow!

Reply on news.ycombinator.com