With the combination of a small number of people + software + servers and robots
We are promoting a new era of company management.
We hope to share part of this process with you in this corner.
October 31, 2022
Before the Corona disaster, we had a cooking club and often made sweets during work.
Usually the interns got involved, and I now have good memories of being roasted gingko nuts on the roof, roasting green coffee beans, and kneading flour to make cookies.
But times have passed and circumstances have changed, which means that we have busted out the cooking supplies.
Iron pans, small wooden plates, ceramic platters, spoons, chopsticks, teacups, mugs, decorative glass plates, tortilla machines.
I remember the party we had when we got oysters directly from Atsukishi with this, and how we went to the trouble of ordering corn flour and made tacos with it, and how we piled up croquettes in a heap with it, and all the things I miss.
While a large amount of cooking supplies were gone, the DIY shelves that had been on display were also dismantled.
We are half sad and half refreshed, and a new project is beginning to run.
October 24, 2022
At the beginning of a new term, we begin to investigate the level of accuracy of the forecasts we came up with just one year ago.
As I talk and write here and there, I carefully collect them one by one and see what has been misaligned. Strangely enough, the gap between expectation and reality has lessened as the years have passed.
Of course, this is because we have found and addressed the factors that caused the discrepancies, but I think the biggest reason is that we now have a gradation of expectations.
We estimate at a 90% predictability level for projects that we can control, at a 60% predictability level for those that we work with close outsiders, and at a 15% predictability level for completely new business plans.
The degree of predictability is not vague, but is determined by synthesizing the extent to which the elements are comprehensively identified, the range of variation for each element, the degree of interaction among the elements, the order in which they affect each other, and whether the intensity of the elements themselves will increase or decrease.
Once the degree of predictability is determined, we can then create a formula to express it and let the computer do the simulation, or use graphic symbols to decipher the flow.
Take the example of a project to launch a new service.
There is one project manager, one designer, and one engineer. Of the three, the designer is inexperienced and wants to use an unknown technology for the new service, and the projection level is 80%, as he wants to create the new service in approximately 4 months.
The unpredictable 20% is mainly due to the level of growth of the members, the availability of unknown technology and alternative options, how the concept is put together, and how it fits in with other work, so if we tighten the project at key points to keep the variability there in an appropriate range, we can probably do it.
That said, it's no trick to only make things highly predictable, so I always weave in less predictable things and chase after the moments that flash before my eyes.
October 11, 2022
If you are alone, or if a problem occurs in a team, it is difficult to identify the cause of the problem.
Emotionally, you could just say, "It's his fault," morally, you could say, "One for all. It's everyone's fault," and data-wise, we can opine, "There's a bottleneck in the workflow.
In response to each, "someone says sorry, we talk about trusting each other more, we try to devise a number check," but what if it still doesn't work?
I quote a straightforward expression from a book on causation.
oooooooooooooooooooooooooooooooooooooooooooooo
The Science of Causal Reasoning "Why? How to Answer the Question "Why?
Judia Pearl (Author), Dana McKenzie (Author), Yutaka Matsuo (Supervisor), Dai Natsume (Translator)
h ttps:// books.bunshun.jp/ud/book/num/9784163915968
For these problems, we immediately come up with the solution of dividing the numbers into a finite and tractable number of categories.
While there is nothing wrong with this solution in principle, it would make category setting somewhat arbitrary.
Also, if the number of variables to be adjusted is greater than a certain number, the number of categories could increase exponentially.
This would make this approach practically inapplicable.
Moreover, in this case, it is possible that not a single element exists in many of the categories.
In that case, the probability would not be estimated at all.
This problem has been called the "curse of dimensionality," and statisticians have come up with some excellent ways to deal with it.
ooooooooooooooooooooooooooooooooooooooooooooo!
It means that factors that we have not come up with in our past experiences, sensations, and conventions are secretly lying dormant and causing bad behavior, and of course, there are most likely multiple factors in a chain of events.
And then, how do you solve it? Unfortunately, it seems that the only way is to take a good amount of time and painstakingly and patiently consider the situation.
From another quote about how it is
ooooooooooooooooooooooooooooooooooooooooooooo!
It's quite simple, Miss Watson.
Sherlock Holmes would say this, and then, as we all know, he would usually surprise his trusted assistant, Watson, with a deduction that was not very easy.
abbreviation
Holmes has other famous lines.
From that, we can see how he thinks.
Here's the line.
When you eliminate the improbable one by one, what remains at the end, no matter how improbable it may seem, is the truth."
ooooooooooooooooooooooooooooooooooooooooooooo!
We live in a world of Corona disaster, yen depreciation, inflation, and more knots that cannot be untied normally.
October 03, 2022
Every year in October, they start all together to prepare for the end of the year.
In particular, the Nenga app project and the New Year's calendar project compete with each other, but this year the atmosphere is somewhat different.
Compared to the New Year's greeting card application, which started smoothly with the new interns meeting each other, the calendar project seemed to be completely stuck in the scene.
That's because the 10th anniversary of LearnO is coming to a climax, and everything is packed into the months of October and November.
We can't force it, because we can't force it to be good.
I think we can give up on the New Year's calendar."
And if you help him out...
My mother enjoys using it.
I would like to see the tree on the stand next year.
I wish we could publish more than one page a month."
And from here and there, we keep coming back to the premise of creating a calendar.
It must be really hard to make a calendar from scratch, but Mogic's love of calendars is extraordinary.
Now let's see if we can do it and if we will have any luck.