Our stopped clock technology is still in its infancy, but it's already reached an accuracy rate of two or more times per day, and there's no reason for us to believe that won't improve dramatically in the future
Lately I've noticed an almost cult-like mentality among stopped clock deniers. And if you had asked me six hours ago, I would've agreed with them—that clock was literally as wrong as it can possibly be. But if you're not looking at that clock right this moment,
abadidea (@[email protected])

if anyone wants to laugh at AI, a friend of mine googled "new york time" and got an "AI summary" claiming 7:14 pm (wildly wrong, it was about 2 pm) and then 40 minutes later I tried and also got 7:14 pm

Infosec Exchange
@joe ah yes, from the famous country song “it’s 7:14 o’clock somewhere”.
@joe it’s always 7:14pm somewhere
@joe on November third it was correct three times in one day! Progress!
@joe I'm drinking jet fuel over here
@joe it is true that the technology is in its early days, and nobody has found a practical application for solving word problems in one relation monoids up to length 11 yet. But once we scale it up to length 12 or 13, the skeptics are going to look foolish for not jumping on the bandwagon sooner
@slava @joe Backpropagation should be able to solve diffeomorphism of 4-manifolds.
@joe A stopped clock is right twice a day, but one that runs backwards is right *four* times a day. Our best scientists are working on running the clock backwards even faster to allow for more frequent accuracy!
@modulusshift Besides, if you're still judging the clock's accuracy by looking at the face, you're still stuck in 2021. We have agents now that will automatically reread the face until the result is accurate
@joe Our bespoke AI-designed model has multiple fingers per hand to increase the amount of times you can read at once! Accuracy rate increasing by leaps and bounds!

@joe lol love this take..

Ppl continuing to be amused by fancy markov chain of grammatically correct and overly confident word strings is everything that’s wrong with this world.

What of actual meaning, and structure of understanding? But I suppose if ppl lack those in and of themselves it’s easier to not see when those are missing (ahem VCs and 99% of SV ahem)

@joe im pretty convinced the stopped clock is sentient
@dotstdy @joe How else could it know the time?!
@joe I wonder if anyone has considered combining readings from multiple stopped clocks with ensemble learning for higher quality results
@bnut @joe our lab is currently testing transfer learning to use stopped clock readings to screen for cancer. We've achieved 96% accuracy on all tests we're willing to publish!
@joe our clock that loses 1 second per hour is aiming to one-up you by boing locally right multiple times a day
@joe we're leaving the typo in because its funnier that way
@joe "But does it scale?" is what everybody wants to know. And yes, yes it does. Buying more clocks increases accuracy almost linearly.
@TomF @joe if we produce 86400 stopped clocks, we'll achieve 100% accuracy

@joe

AND WITH ZERO PESKY LABOR COSTS

@joe @simrob [first day of standard time] wow our accuracy is trending WAY up
@joe a stopped 24-hour clock is only right once a day
@AlgoCompSynth sounds like you need to upgrade to Stopped Clock 16 Pro to enjoy the full benefits of stopped clock technology
@AlgoCompSynth @joe That's why I will create competition! I intend to produce an innovative stopped clock! An 1-hour stopped clock! It will be right 24 times a day! Pre-order now for just 500$!
If we extrapolate that in the other direction, a stopped 6 hour clock would be right for 4 times per day.
@joe I have improved the design by rotating it counter clockwise at a rate of one rotation of the long hand per hour and now I’ve duplicated the precision: the clock is right four times a day.
@MyLittleMetroid @joe That’s all very well in raw numbers, but our customer surveys say that consumers really care about performance per watt, a metric in which stopped clocks have a near infinite advantage.
@joe OMG when did you join mastodon
@joe Before we smashed the clock mechanism with a hammer yesterday, it was accurate 0 times a day. If you look at the graph, we can extrapolate that by the end of next week...
@joe check out my new AI that can tell the exact time with only a photo of a stopped clock !
@joe @voltagex I'm convinced it's correct at least somewhere around 38 times a day, if not more... From a certain point of view.
@petelawler @joe @voltagex Exactly! Whenever the clock is wrong, it's because you're reading it wrong. Learn how to read the stopped clock more effectively. Online courses available at a 20% discount only today!
@joe a lot of negativity in this space but we're here to build and people should respect that

@joe

We doubled accuracy by moving from 24h time to 12h time. We feel further improvements can be made by moving to 8h, 6h, and 4h time...

@joe beside those two ticks, the rest of the day still makes it a valid unreal time clock.
Mixing real time and unreal time technology is interesting.
@joe if we pour trillions into stopped clock production we can one day have enough stopped clocks that for any given minute of the day there will be at least one stopped clock in every home that’s guaranteed to be right

@joe it turns out if you hire a person to move the clock's hands every hour you can get the stopped clock pretty close to the correct time

We pay them a few cents per clock moved. We expect to receive about ten million dollars for our first series A and-

@joe If we introduce active cooling, I believe we can overclock the stopped clock for improved performance.
@joe @nazgul Pull the hour hand off and it’s right 24 times a day. #LessIsMore
@xvf17 @joe @nazgul You’re not thinking big enough - leave only the second hand for both better precision AND accuracy 1,440 times a day!
@whybird @xvf17 @joe @nazgul Masterclass application of the philosophy behind #YAGNI You ain't gonna need it
@joe we need to harvest all the stopped clocks in the world. as everyone knows, a huge amount of shit data is much better than one correct data point!

@joe Our Intelligent Time service has over ten thousand stopped clocks and a sophisticated surveillance system closely examining what lots of people are looking at from moment to moment.

When we report the time based on the stopped clocks that we see the most people looking at, we find that far more than twice a day our time is accurate, or at least plausible enough that most of our users don't notice.

Once we can command even more stopped clocks and more surveillance, it'll be amazing.

@joe I badly need a sticker of this toot

@joe well i've heard enough. time to replace all the old clocks we use with these new ones

one question though, is the smell of the earth burning coming from the stopped clocks or what?

@joe This just in: investigation of the recent 80% accuracy claims of SCT finds that Stopped Clock Technology firms have acchieved that number by paying hundreds of thousands of people in colonized nations to manually stop clocks at the correct times and return the results. SCT proponants responded for comment, saying this is just a temporary part of the process to train the clocks to stop correctly on their own
@joe Stopped Clock Tecnology stock prices soar as new advancements are made. previously the accuracy of stopped clocks were doubled when researchers moved from using 24-hour clocks to 12-hour ones. After a period of stagnation they accieved the same feat again with the brilliant innovation of 6-hour clocks.

@joe but do you have a "moat" to satisfy investors of infinite returns in perpetuity?

What would that moat be? Keeping all the world's clock's ticking so as not to accidentally steal your IP?

@joe

We have applied for a 1.7 billion grant to investigate "SSC risks", sentient stopclocks that choose deliberately to tell the wrong time.

@joe indeed. Multiple stopped clocks, working in parallel, could be used to accurately map all the seconds in a day. Smaller hands could give us access to the milliseconds, and so on.
@qalthos your display name really is ahead of the curve!