Chaotic Thoughts

Miyazaki's Love Letter

logrusadm@gmail.com (Anton Golubtsov) — Sun, 29 Mar 2026 22:02:32 -0700

I was walking my dog when the thought arrived — that Hayao Miyazaki’s films are a love letter to ordinary women.

In Nausicaä, we see a little girl who believes in the friendly intent of nature, standing between her people and a seemingly hostile world, establishing peace through sacrificing herself for the things she loves. Similarly, in Spirited Away, Chihiro takes care of her family and her friends, starting as a person who doesn’t care much about anything but herself. She is a child, after all — moving from one city to another, new school and so on. Why would you expect any different? But then she grows through hard labor and a bunch of obstacles. She demonstrates care for people you wouldn’t expect her to care about — Lin, the girl she met in the bathhouse who became her friend, then even Yubaba’s baby, then Haku. They were all a little hostile towards her at first, and still that pulling desire to help, to save the day.

The Purpose of a System Is What It Does

logrusadm@gmail.com (Anton Golubtsov) — Sun, 28 Dec 2025 21:37:36 +0300

Examples of applying the POSIWID lens.

Introduction

In management cybernetics, there is a well-known formulation: “the purpose of a system is what it does” (POSIWID). It is commonly used as a practical way to analyze complex systems—organizations, institutions, social practices, and recurring patterns of behavior.

The idea is to set aside declared goals, missions, and explanations, and instead observe the actions and effects a system reliably produces over time. Not isolated decisions or stated intentions, but stable patterns—especially those that become visible under stress, growth, or conflicting incentives.

One Year, 2025, in Books

logrusadm@gmail.com (Anton Golubtsov) — Wed, 24 Dec 2025 08:28:38 +0300

Intro

Another year is coming to an end. People summarize their year and make plans. Reflecting on mine, I once again noticed that books — like many other things we choose and spend time on — tend to reflect who we are at a particular moment in life rather accurately.

This year, that mirror suggests I was more concerned with people and personal development than with technology. Perhaps that’s self-deception, or a way of avoiding other topics. It doesn’t really matter. Books are just books.

Hierarchies Are Wonderful

logrusadm@gmail.com (Anton Golubtsov) — Wed, 02 Jul 2025 22:57:28 +0300

Recently, I’ve often heard that hierarchies are bad, supposedly representing nothing but inequality and oppression. This topic unexpectedly surfaces in seemingly unrelated contexts. I encountered it in a book about startup culture, my children brought it up from school, and there was tension in chats on unrelated topics whenever structures resembling hierarchies were mentioned. On one hand, I agree that some hierarchies can be built on force and suppression. On the other hand, I see them more as symbiotic structures rather than tools for power struggles.

Heavenly Doors and Flaming Swords

logrusadm@gmail.com (Anton Golubtsov) — Wed, 02 Jul 2025 22:53:33 +0300

There are people who feel like a flaming sword — the kind Peterson writes about in “We Who Wrestle with God,” burning away everything unworthy just by standing near them. We avert our eyes, or we try to dim their light.

Analyzing my own reactions to such people I found envy, a reminder of failures, or weaknesses or a sense of a vast gap in the strength of the spirit. All of this is accompanied by a feeling of internal constriction and resentment, although it could be associated with the aspiration to achieve more and become better. This however requires accepting one’s own imperfections, which is valuable in itself, but also makes being near the ideal less scorching. Perhaps if you have never felt this way, you are already pure enough inside.

The Impact of Mental Fatigue, Task Monotony, and Data Skewness on Data Annotator Performance

logrusadm@gmail.com (Anton Golubtsov) — Fri, 11 Apr 2025 21:12:19 -0700

Intro

There are two topic that people rarely discuss when in comes to data annotation. First, that annotators are not just things that needs to be trained, and whose performance needs to be closely monitored but regular human beings like any one of ourselves as people who can get tried, or distracted. And second, how the data we push through those people influences their annotators’ performance.

As I feel compassion, and deep respect for annotators I work with on daily basis I wanted to cover that topic in one of my writing but never had time to do so as well as bandwidth to conduct a proper study. With help from, AI and all researches who has done the ground work, I can at least share a short summary.

Review: The Geek Way by Andrew McAfee

logrusadm@gmail.com (Anton Golubtsov) — Mon, 09 Dec 2024 19:52:18 -0800

The book focuses and explores in depth the key traits of successful companies like Amazon, Apple, Microsoft etc. It differentiates four main areas: ownership, openness, speed, and science. From all four openness is mentioned as the key to all other features as it naturally enables them.

For myself, I summarize the book as two aspects: environment, and speed.

The company needs to focus on creation of an environment where ideas can move freely, basic needs like access to code (including changing the code), data, documentation, support easily available. In addition to that negative feedback not only welcomed but actively sought so any issues can be discovered as soon as possible so they do not become too large of a problem. This requires an atmosphere where people are not punished for mistakes and so there is no implicit incentive for covers up.

Bright Cocaine: colors and dopamine

logrusadm@gmail.com (Anton Golubtsov) — Thu, 05 Dec 2024 20:47:17 -0800

Intro

As continuation of my attempts to regain control over my attention, two month ago I switched my phone to first black, and white mode, and then I applied the same filter at 30% on my laptops and then reduced the filter intensity to 20-30% on all my devices. I helped a bit, actually black, and white helps quite significantly but it is a bit tiering to use.

So after the first month I’ve noticed the following changes:

Python Pipes

logrusadm@gmail.com (Anton Golubtsov) — Fri, 29 Nov 2024 19:38:26 -0700

I’ve always wanted to have a way to build data processing pipelines in Python using pipes, like this range(10) | F(is_odd) | P(lambda x: x * 2), instead of functions and generators and maps and loops. So I’ve tried …

The idea is pretty simple: let’s create a class with implemented OR and ROR operators, the pipes.

    def __or__(self, other):
        other.source = self
        return other

    def __ror__(self, other):
        self.source = (
            iter(other)
            if not isinstance(other, (str, bytes)) and hasattr(other, "__iter__")
            else other
        )
        return self

The tricky part was implementation of __next__ since I wanted it to be a lazy operation. After a few trials and errors I’ve ended up with a pretty simple approach where the wrapping class implementing the pipe will call next to its source, added by OR or ROR, apply a transformation and then return the result of the transformation.

Experimental Power Demo: Frequentist vs. Bayesian Power Visualization

logrusadm@gmail.com (Anton Golubtsov) — Thu, 31 Oct 2024 17:51:01 -0700

a simple visualization of the difference between frequentist and Bayesian power and how effect size, noise (standard deviation), and sample size affect the results.

All Things You Hear

logrusadm@gmail.com (Anton Golubtsov) — Wed, 16 Oct 2024 22:05:37 -0700

I like the idea I got from one of Joscha Bach posts that everything we hear is automatically executed by our brain like it is own thoughts even when we don’t pay attention to it. And that it is a security vulnerability that it opens a gate for all sorts of exploitations. I keep returning to this thought frequently and so I decide to put it on “paper”.

The summary of my thoughts on the topic. We have that loop hole that can be exploit in multiple ways from little verbal abuse, or manipulation to full-fledged propaganda attacks. The most devious part is that it can crawl under your skin even if you don’t even pay attention to it. It is probably even more contagious when you are not pain attention since your guard is off. You can resist certain ideas when you hear then once, twice, ten times but after a hundred of repetitions they will eventually get through unless there is something in you that will make those ideas completely unacceptable for you.

Recommendation System on HNSW and Exponential Moving Averages

logrusadm@gmail.com (Anton Golubtsov) — Fri, 11 Oct 2024 21:29:10 -0700

Intro

I was reading the original paper on “Hierarchical Navigable Small Worlds (HNSW)” https://arxiv.org/abs/1603.09320 which I found much easier to understand than all those YouTube videos I tried to watch and articles to read. HNSW is a probabilistic data structure for searching neighbors in multi-dimensional space. One of practical applications is search of semantically close objects. Reading that paper and some other activities made me curious if I can quickly implement a recommendation system which combines three things: HNSW, moving averages, and randomness.

Moving Averages

logrusadm@gmail.com (Anton Golubtsov) — Wed, 09 Oct 2024 20:49:48 -0700

I was curious about use of averaged vector embedding for recommendation purposes, and then I started wondering if instead of averaging I should try other metrics like median or top percentiles to focus on more frequent scenarios and reduce the influence of outliers.

And then the question was: imaging that you want to use it in production, how can you compute averaged embedding for millions of users ideally with instant updates and without offline data processing in bulk.

Chaotic Good

logrusadm@gmail.com (Anton Golubtsov) — Mon, 07 Oct 2024 21:34:24 -0700

A few days ago a manager from a sister team asked me why I have “Chaotic good” in that field “What I do” of my work profile. The question now occupies my mind so I need to drain it into something.

In short, it is a silly meme from a few years ago as I decided to use as a work motto instead of “stupidity and courage” as I fell that the new one better reflects the type of work was doing at the time.

Classic Salad Dressing

logrusadm@gmail.com (Anton Golubtsov) — Thu, 03 Oct 2024 18:20:19 -0700

Ingredients

I have no clue how much of ingredients you will need all measurements are approximate.

Balsamic vinegar, or any acidic substance. Try lime juice, it has nice aroma touch. 1 or 2 tbs. Probably.
Olive Oil or any oil try peanut oil. 3 or 6 tbs. Who knows how much.
Mustard. This thing stabilizes to keep the emulsion. A bit of it.
Salt. A pinch?
Pepper. A bit.

The process

Mix everything together and stir for a minute till the emulsion built up.

Data Sharding(Partitioning) Algorithms

logrusadm@gmail.com (Anton Golubtsov) — Mon, 30 Sep 2024 17:20:24 -0700

I used to work close with incredibly smart people who was dealing with things like data sharding on daily basis from them I learned a lot on that topic. Later I moved to a different role where that knowledge was not needed and faded away over the time. Here I’m trying to reclaim to myself that long forgotten knowledge.

Intro

Sharding is a process of assigning an item to a shard - a smaller chunk of data out of a large database or other service. The general idea is that we can distribute data or service across multiple locations and handle large volumes of data or handle more requests and with replication we can scale even more and make the system more resilient etc. But we need to have clear rules on how we assign partitions aka shards so that we can route requests to the right location.

Interview Question: logs parsing library

logrusadm@gmail.com (Anton Golubtsov) — Sun, 29 Sep 2024 17:30:06 -0700

One of the question I often ask in my interview is to design a log processing library:

You need to write a library for processing logs in the following format:

timestamp<TAB>message

The library will be handed over to a different team for further maintenance and improvements and so maintainability and expandability is the most most important requirements.

The library need to support the following operations out of the box:

filtering
counting
histograms

The original version also included some language and background specific expectations I never include in my assessment because I feel that they put the candidate into a position when they need to read my mind to meet those expectations.

An Interview Question: Optimization of Disk Read costs

logrusadm@gmail.com (Anton Golubtsov) — Tue, 24 Sep 2024 17:27:24 -0700

One of the questions are really love asking during coding interviews is this:

Given a continuous stream of words, a dictionary on disk and cost associated to read from disk, create a stream processor that returns true when a word exists in the dictionary while minimizing the cost of reading from disk. Example:

Dictionary: {Dog, Cat, Bird, Lion, ...}
Input: [Dog, Cat, Aghd, ...]
Output: [True, True, False, ...]

The output is true, true, false because dog and cat exist in the dictionary of words while Aghd is not considered a word.

Writing User Stories and Requirements

logrusadm@gmail.com (Anton Golubtsov) — Wed, 18 Sep 2024 09:03:53 -0700

Recently I was reading though a bunch of technical designs and I’ve noticed a common mistake when it comes to writing user-stories and requirements - assuming a solution. The biggest issue for me when I write requirements myself is that than whenever I include a part of a solution I’m thinking about into the requirements it limits my ability to innovate since I’m bound to a specific solution. In many cases I observed improvements in my designs when I was focused on what the customer needs rather on fulfilling requirement tied to my first and probably not a bright idea.

High Tech Energy (Attention) Vampires

logrusadm@gmail.com (Anton Golubtsov) — Sun, 15 Sep 2024 15:21:08 -0700

It is interesting to observe that any endeavor where attention is one of key metrics or key drivers. Regardless of the company size end up in the same hell pit of attention craving and optimization for it. Even small single person blogs that teach us to be a better person, engineer, or somethings are prone to that. Many of them, those I used, slowly became “Energy Vampires” to me constantly seeking for my attention.

Cache eviction policies: LRU vs random vs p2c

logrusadm@gmail.com (Anton Golubtsov) — Sun, 15 Sep 2024 10:22:58 -0700

Every so often I interview senior software engineers for Amazon. Where I ask more or less same questions in each interview. One of them requires adding a caching logic to get better results. I’ve noticed that the interviewee make one of to mistakes that blocks them from standing out as a software engineers:

they don’t know, or talk about conditions under which a cache will do the best. Primarily, how a request frequency distribution affects cache performance.
they don’t know the standard library of a programming language of their choice.

Here, we will try to address those issues.

JTBD: day one - how to make a product

logrusadm@gmail.com (Anton Golubtsov) — Sun, 25 Aug 2024 11:22:58 -0700

bare minimum about the Jobs To Be Done framework

JTBD: why people buy

logrusadm@gmail.com (Anton Golubtsov) — Sun, 25 Aug 2024 11:22:58 -0700

Emotions

All problems comes from some emotions like frustrations, anger, sadness, guilt
Emotions help to think to highlight a problem, to move you attention towards something, so you can fix it.
Emotions give energy to act.
90% of decisions are automatic

Internal narrator and conflict resolution

Internal narrator is a conflict resolution mechanism or a consensus protocol. It stitches the pieces of really when something is done in a conflict and explains why it is ok. “I decided not to drink BUT today is ok because it was a tough day.”
- This reminds how LLM works they predict w
- Brain predict future and uses senses to correct its predictions. When something is not going at it was predicted it generates signals to correct it though “pain”, pleasure OR lies. It reminds me of LLMs with their hallucinations and predictions of next character they need to produce.
Lisa Feldman Barrett wrote a book How Emotions Are Made where she explains how brain constantly hallucinate of future, lies, and use emotions to correct human behavior.

Internal investor

The brain acts like an investor, it observes the available options and tries to invest effort/resources in things which more likely will give higher-return.
The brain tries to reduce costs which is associated with “pain” centers in the brain and increase the reward which is associated with “pain”. See: Neural Predictor of Purchases Purchase == (pleasure from purchase - “pain” of losing money) > 0
In the product management we have the same process persuasion of pleasure (positive emotions) and “pain” avoidance (negative emotions).
- It will be interesting to see if these mechanism can be used to improve kids performance at school by tweaking how much “pain” they get to receive a reward. Like doing less for the same reward, or doing the same amount but for higher reward.
- It is also interesting is there some level of “pain” which practically block the path forward. Could it be that kids are not doing something till the very last moment because only at the very last moment the negative impact of not doing something finally lower than effort of doing something.
- Or how can we use it in other aspects of our life?

![[Consistency-Theory.png.webp]]

JTBD: day two - Advanced Jobs to Be Done

logrusadm@gmail.com (Anton Golubtsov) — Sat, 24 Aug 2024 11:22:58 -0700

Key insights

jobs are more important than product of features because they help to find a product or features but not the other way around
head of the customer is the main source of truth, not the head of a product manager
analyze customer jobs and job graphs, not a product.

Sequential jobs

step-by-step
the value is received in the end of the entire job chain
because a next job depends on the results of a previous job
you need to know dependencies and possible branches
fix issues first because they affect the whole chain
the chain issues, or results affect high-level jobs

Frequent jobs

viral jobs - jobs which the user does for somebody or with somebody else like “prepare slides so they can present them in a meeting,” or “write a restaurant review”, “work together on slides”

Tax/Penalty jobs

When something goes wrong the user needs to adjust their plans those adjustments create “penalty” or “tax” jobs. For example, the user arrives to a gas station but they couldn’t pay because something was wrong with credit card payments and they had no cash. In this situation the user invested their resources (time, and effort) to get to the gas station, and now the user needs to invest more resources to get what they need. The brain does not like mis-calculations and the user will remember the pain associated with a situation or a product. Next time the brain may try to avoid the pain associated with your product

JTBD: reading list

logrusadm@gmail.com (Anton Golubtsov) — Sat, 24 Aug 2024 11:22:58 -0700

Strategy and business value

Why people buy

Must read

- How People Change - The Body Keeps the Score - The Molecule of More

JTBD: B2B specifics

logrusadm@gmail.com (Anton Golubtsov) — Fri, 23 Aug 2024 11:22:58 -0700

personal goals go first
new product new challenges
- teach, train, hire, fire
- explore internal kitchen
- showing useful to the potential customer information in advance 3x increases probability of success
personal contact is more important than a product. people perceive familiar people as more reliable so more likely to provide support in the future and so secure the position of a decision maker who buys the product
personal jobs is a sales driver - 90% personal goal of a decision maker (promotion, status, job security) and only 10% business goal
- for decision makes who has higher personal connection to the business goals, like CEOs, the distribution is 50/50 between personal and business goals.
- Personal goal of decision makers
  - job security - personal contact, remember?
  - status - “I’m better at something”, “Look how cool I’m”
  - fun - “I really like that thing”
  - support in the future and risk mitigation - related to job security
- insight -an expensive product can be 8x likely to be bought if it is aligned with the decision maker personal goals
you need to understand that is the decision making process and who is involved in the process so you know who else you need to talk to.
- also try to understand that are the decision making criteria so you can focus on meeting those criteria
types of people in B2B
- Decision makers - talk to them first the rest are less important or sometimes irrelevant
- Budget holders - they can block the decision
  - don’t hesitate to ask about the budget and make sure you understand the process
- Influencers - people inside (or outside) the company who can influence the decision one way or another. Work with then if you can
- Recommenders - people who can be asked for their opinion about your product
- Saboteurs - people who can be strongly against your product - like a decision maker who bought previous solution
talk to decision makers first

What to take out of a B2B interview

jobs and goals both business and personal
problems
decision making process, people involved and their roles, evaluation criteria
budget - first establish contact with your respondent before asking this question - mid or end of an interview can be a good point to ask

How to find an decision maker

work as a team
send some that personalized email like “you were recommended to us by …”
- recommendations from real people work the best
ask for people you can talk to
bring some value with you
LinkedIn is a good place to start

JTBD: day three - solution value and how to create it

logrusadm@gmail.com (Anton Golubtsov) — Fri, 23 Aug 2024 11:22:58 -0700

Solution value can be defined as a difference between perceived benefits provided by a solution and perceived costs associated with a solution. In addition to that there can be a perceived cost of switching to the solution. We use “perceived” here since there is usually a gap between reality and perception and value is in the eye of the beholder.

$value(solution) = benefits(solution) - costs(solution)$
value is relative. It is hard to assign a value number to a solution so people usually compare one vs another.
main paths to add value:
- add benefits
- reduce costs
- both
- sometimes it is cost increase but providing much more benefits
aha-moment or value activation - aha-moment is the moment when the user realizes the value of a solution. For example, for a new expensive car the status value of a car can be activated at the point when you gave a ride to a friend and they said “wow, that was great.” For a report, it can be a moment when you receive positive feedback for the quality of the report and not the moment when you generated a report.

Top mechanisms to bring more value

do an important job nobody else does
fix problems in the main job
provide base quality for your solution - people expect it as bare minimum
eliminate low-level jobs
do more jobs within a user using one solution so the user uses your solution more frequently and value it more as a tool
move to a different segment with a same solution
do the immediate next job after current job - Figma is a good example, current job “create a report”, and next job “share report”
reduce the cost of a job
add more benefits to a job
fulfill basic needs: status, safety, social contact etc
communicate that a solution fulfills one of the basic needs
wake up a sleeping job - higher level jobs which can be completed more efficiently but using a different solution
correct expectations - the user may have unreasonable expectations correcting them in advance will reduce disappointment when unreasonable expectations are not met

An algorithm for solving business problems with a product

clarify and challenge the goal
find mechanisms which may help to achieve the goal
find raw material to pick right mechanisms: analytics + JTBD interviews
apply mechanisms: Airtable - How to pick value mechanisms
rank mechanisms by RICE (Reach, Impact, Confidence, Effort)

On RICE

RICE (Reach, Impact, Confidence, Effort) is a mechanisms for ranking different solutions. Confidence can be tricky to estimate you can try to use Idea Prioritization With ICE and The Confidence Meter for estimations.

JTBD: B2B specifics

logrusadm@gmail.com (Anton Golubtsov) — Thu, 22 Aug 2024 11:22:58 -0700

Before an interview

in a customer interview studying the reality not their imagination.
you should have a goal before an interview so:
- come prepared
- have a hypothesis
- know your customer segment unless you are looking for one
preparation steps:
- research the topic if you are not familiar with it - it will help you to earn the interviewee’s trust and so get more insights out of the interview
  - know the domain specific language
  - talk to experts if needed
- decide what the duration of the interview should be. It is usually between 30-90 minutes. 30 minutes maybe too short since you may need to earn some trust first or you may want to go deeper into the details. 90 minutes may be mentally and physically hard.
- prepare an interview script but remember that it is a blueprint and you can adjust it as you go
recording and notes
- video recording is nice to have if you want to review but it is better to use notes so you don’t need to spend another hour which you can use for an interview to view the recording
- ask for permission to record a video even if you asked it before the interview
- inform that you will be taking notes
- avoid any distraction it is important for establishing the contact and trust
- video is preferred since 55% of information is non-verbal
- prepare a notes template to fill in
- you need to feel comfortable during the interview: meditate, find a place from which you will be comfortable to interview

During an interview interview

Contact

smile
open posture
look into the camera or keep eye contact
use active listening - it help to reduce the stress through affirmation

Set boundaries

introduction. try to switch to less formal language (for languages with different formality modalities)
explain the interview process
- in B2C hide product details till the last moment or do not expose them at all (also you may not have a product yet)
- explain the timing
tell that you will be taking notes
tell that you may ask about their feelings and emotions
let them know that they don’t have to answer questions they don’t want to answer
ask if they are ok with video recording

Interview

don’t reveal the goal of your interview to avoid lies and hallucinations in an attempt to please you.
start with simple and qualification questions to give a sense of safe space and to know better qualification factors our like their interests or life priorities etc. Those factors can be very insightful.
avoid domain specific terms and slang - use general language
don’t ask about the future or imaginary situations - people will lie and hallucinate
ask about past experience and to tell a story
- if a respondent drifts towards imaginary situations carefully redirect them to the main topic - the past experience
ask open questions
ask problems and associated emotions and their scale
- be careful with wording not to shut them down by rejection or misbelieve
one question at a time
avoid interrupting, carefully correct the topic of discussion
talk less, don’t fear the awkward pauses. The pauses can nudge the respondent to start talking about things they wouldn’t tell otherwise. People don’t like silence and will try to fill it with their thoughts.
ask follow up questions like “why?”, “so what?” etc
if a team want to join
- limit the number of participants
- come up with a protocol for asking questions by the team members
- introduce the team
- you may want to put team of camera if they are not actively talking

After the interview

Analyze the interview as soon as possible while you have all the context fresh.

JTBD: day four - segmentation

logrusadm@gmail.com (Anton Golubtsov) — Thu, 22 Aug 2024 11:22:58 -0700

Choosing a segment is one of the most important decisions. It will define the results you can achieve.

We can’t build a product for everyone.

A free interpretation of from Competitive Advantage: Creating and Sustaining Superior Performance: Porter, Michael E.: 9780684841465: Amazon.com: Books

Our goal is to become the best in a specific, and maybe in initially narrow segment, through precise knowledge of the jobs in the segment and implementation of a significantly better solution that competing solutions.

JTBD: day five - conversion and sleeping jobs

logrusadm@gmail.com (Anton Golubtsov) — Wed, 21 Aug 2024 11:22:58 -0700

What does the customer purchase depend on

#	Dependency	What to do
1	attracted a lead with a job we have a solution for	Find, and choose a right segment. JTBD for a segment selection + marketing
2	offered higher value than current solution or competitors	Create value by utilizing JTBD for product discovery and comparing our solution, and the competitors
3	the lead realized/understood the value of our solution	Iterate over hypothesis for loading the value into the head of the lead and test them in decision interviews
4	the lead does not have concerns, or fears toward our solution	Iterate over hypothesis for decision barriers elimination and test them in decision interviews
5	the lead does not have concerns, or fears toward the job execution	Iterate over hypothesis for job completion barriers elimination and test them in decision interviews
6	the lead was able to activate the value: didn’t abandoned the solution due to high effort/costs and realized what value it will receive	*Discovering value, finding aha-moment* and moving the value communication earlier in the funnel,** test them in decision interviews and UX-tests
7	the lead was able to get through	Iterate over decision interviews and UX-tests to fix problems

#tbd decision interviews

JTBD: days six and seven - Product Strategy

logrusadm@gmail.com (Anton Golubtsov) — Tue, 20 Aug 2024 11:22:58 -0700

Three insights

Product strategy is not a kind of magic; it is an algorithm.
Strategy is a path, not a destination.
Small daily decisions we make are a part of the strategy.

Note: Companies that operate outside of market rules do not need a strategy since their decision-making process is not market-driven—for example, the government.

The main mistake is the ignorance of current trends.

Sandy san, Our current success is the best reason to change. We must be permanently dissatisfied in order to succeed.

Ultra-locality in Decision Making and Free Will

logrusadm@gmail.com (Anton Golubtsov) — Sat, 23 Mar 2024 21:15:45 -0700

This time we explore the wonderful world of ultra-locality in decision-making and its connection to free will, good, evil, and God.

Part One: Ultra-locality and Free Will

The same Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans | Lex Fridman Podcast #392 podcast I mentioned in the previous post AI, people, trees, and mushrooms: the same software different hardware triggered another chain of thought. Joscha was talking about how our neurons always operate using data available right here and right now. That is enough to build complex systems like the human brain. Working together, neurons form parts responsible for memories, image processing, data buses, etc. But ultimately, each of them individually works only with data provided by other neurons. In a similar fashion, neural networks in GPTs are just a multiplication of matrices connected with each other, forming memories, attention, generation, etc.

AI, people, trees, and mushrooms: the same software different hardware

logrusadm@gmail.com (Anton Golubtsov) — Sat, 16 Mar 2024 21:06:45 -0700

Exploring the idea that all living things have spirit or the ability to run neurological signals.

Recently, I listened to Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans | Lex Fridman Podcast #392 where Joscha and Lex discussed different ideas about consciousness, neurology, and AI. At one point, they talked about the ability of all types of cells to process neurological signals. The key difference is that neurons can process data much faster, over longer distances, and interact with more neighbors at once.

The Matrix: A Simulation, a Game, Reincarnation, or Hallucination?

logrusadm@gmail.com (Anton Golubtsov) — Sat, 09 Mar 2024 17:38:57 -0700

For quite some time, a few thoughts have been haunting me. What if all we see doesn’t exist, and we are all hostages or participants of a game.

The Beginning: The Matrix and Rick and Morty

It all started a while ago. In high school, I was into cyberpunk, reading and watching about hackers, virtual reality, etc. Neuromancer, Johnny Mnemonic by William Gibson, Labyrinth of Reflections by Sergei Lukyanenko, The Matrix, and The Lawnmower Man were my go-to entertainment. Then I forgot about it until the recent generative AI explosion and… a few episodes of Rick and Morty. In one episode, Morty plays a VR game where he starts as a newborn with no memories of life outside the game and lives an entire life until he dies at 60-80 years old. In another episode, he gets stuck in a game where his consciousness is fragmented into pieces, acting as an entire world of independent agents.

Lean development, customers and how it is connected to Amazon writing culture

logrusadm@gmail.com (Anton Golubtsov) — Tue, 05 Mar 2024 17:23:16 -0700

Amazon is famous for its writing culture which I discovered later in my career. The more I wrote the easier it was to apply the similar approach to other aspects of software development.

Introduction

When I joined Amazon, I transitioned from a company with a markedly different culture, particularly in writing and development processes. Initially, my feelings toward Amazon’s writing-centric culture were skeptical. However, over the next seven years, I gradually embraced and excelled in the Amazon style of writing.

Want to be connected? Think twice.

logrusadm@gmail.com (Anton Golubtsov) — Sat, 02 Mar 2024 13:19:20 -0700

In the last few years companies try to connect everything to the internet and make everything be software defined from door bells, and photo cameras to cars. There are a few bad things about it.

These days every teeny-tiny piece of technology tries to be connected to the internet. This is usually sold by large, and not some large companies as some sort of convince so you can control everything from everywhere. The approach has its pluses but it also has its minuses. The shiny technological all connected future does not seem to me as appealing as we are told.

い (i/e) and な(na) - adjectives in Japanese

logrusadm@gmail.com (Anton Golubtsov) — Sun, 18 Feb 2024 23:15:20 -0700

There are two types of adjectives i and na and …. some peculiarities.

I’m not an expert in Japanese phonetics just love to learn new things like …. you know Japanese. So it is just how I perceive the language. Also everything is simplified to the level when it can be understood by everyone.

There are two types of adjectives in Japanese: i-adjectives and na-adjectives. The types are named by the sound which connects an adjective and a noun. These two types are interesting because they conjugate differently. Na-ones act more like nouns and i-adjective have their own way.

One Interesting Code Challenge from Leetcode: finding anagrams in a string

logrusadm@gmail.com (Anton Golubtsov) — Sun, 29 May 2022 11:22:58 -0700

In my coding interviews I often use a simplified version of this challenge from Leetcode. The simplified version I use: Given two strings s and p, return true if s contains an anagram for p. I like this task because the solution and be improved little by little and there are a lot things to discuss. From algorithmic complexity to CPU cache level optimization. Here I want to walk you through how I solved that challenge for the first time.

Book review: Fluent Python by Luciano Romalho (Part 2)

logrusadm@gmail.com (Anton Golubtsov) — Mon, 16 May 2022 21:38:13 -0700

The second part of the book focuses on different callables. You will know about type hints for Callables, protocols, closures, decorators and other related things.

About

logrusadm@gmail.com (Anton Golubtsov) — Thu, 12 May 2022 20:50:09 -0700

Hi, I’m Anton and I’m a software engineer. Mostly self taught. I’m a bit chaotic in my hobbies today I do a lot of film photography, tomorrow I’m trying to figure out how to design simple 3D models in Fusion 360. From time to time I have that etching need to write something down and this the place I’m scratching that need.

Besides that I love the idea of open information and don’t really like paywalls so I put everything here under CC BY 4.0

Book review: Fluent Python by Luciano Romalho (Part 1)

logrusadm@gmail.com (Anton Golubtsov) — Mon, 02 May 2022 22:30:35 -0700

I’ve just finished reading the first part of the Fluent Python book and it is so exciting that I decided to write about the largest gems I discovered in the book.