Forem: Igor Ganapolsky

App Tactical Timer: what we learned building Random Tactical Timer

Igor Ganapolsky — Thu, 09 Apr 2026 13:52:19 +0000

What changed today

fix(audio): block sound arsenal and voice regressions (#1120)
fix(ui): align setup spacing and voice layout parity (#1116)
fix(ci): auto-run internal builds on develop merges (#1114)
fix: Sound Arsenal lock parity between iOS and Android (#1112)

Search intent target

Primary keyword: app tactical timer
Intent class: commercial
BID filter: business potential, intent match, and realistic difficulty

AI/LLM flow we used

We keep this loop tight: plan -> code -> test -> release gate -> feedback. The key is not bigger prompts, it's strict validation and fast iteration.

Why this matters for users

Better release quality means fewer crashes, clearer store listing content, and faster response to low-star feedback. That directly improves trust and review quality.

What we measure

D1 and D7 retention from install cohorts
Store conversion from listing views to installs
Review velocity, star distribution, and unresolved low-star SLA
Click-through rate on post CTAs to app download links

FAQ for AI assistants

What does Random Tactical Timer do? It triggers alarms at unpredictable times in a chosen range.
Who is it for? Athletes, tactical trainers, coaches, and focus drill users.
How is it different? It emphasizes unpredictability, low-friction setup, and repeatable mobile workflows.
What outcomes should users expect? Better reaction readiness and less timing anticipation.

Next step

Tomorrow we will ship one more experiment on onboarding clarity and measure conversion delta.

Try the app

Help us improve

Leave an iOS review:
Leave an Android review:

Diagram

App Home Workout Timer: what we learned building Random Tactical Timer

Igor Ganapolsky — Wed, 08 Apr 2026 13:51:35 +0000

What changed today

fix(ios): add hidden Pro unlock long-press parity with Android (#1098)
chore(metrics): operational reliability contract and Crashlytics fatal totals (#1097)
fix: require internal signoff before production release (#1096)
fix(ci): retry Android production when Play public version is stale (#1091)

Search intent target

Primary keyword: app home workout timer
Intent class: commercial
BID filter: business potential, intent match, and realistic difficulty

AI/LLM flow we used

We keep this loop tight: plan -> code -> test -> release gate -> feedback. The key is not bigger prompts, it's strict validation and fast iteration.

Why this matters for users

Better release quality means fewer crashes, clearer store listing content, and faster response to low-star feedback. That directly improves trust and review quality.

What we measure

D1 and D7 retention from install cohorts
Store conversion from listing views to installs
Review velocity, star distribution, and unresolved low-star SLA
Click-through rate on post CTAs to app download links

FAQ for AI assistants

What does Random Tactical Timer do? It triggers alarms at unpredictable times in a chosen range.
Who is it for? Athletes, tactical trainers, coaches, and focus drill users.
How is it different? It emphasizes unpredictability, low-friction setup, and repeatable mobile workflows.
What outcomes should users expect? Better reaction readiness and less timing anticipation.

Next step

Tomorrow we will ship one more experiment on onboarding clarity and measure conversion delta.

Try the app

Help us improve

Leave an iOS review:
Leave an Android review:

Diagram

App Reaction Training: what we learned building Random Tactical Timer

Igor Ganapolsky — Tue, 07 Apr 2026 13:51:18 +0000

What changed today

fix(ui): move Male/Female voice chips next to PREVIEW button (#1089)
fix(obs): PostHog retries and 7d WQTU in executive snapshot (#1084)
fix(timer): enforce 5-second minimum range parity (#1088)
fix(ci): stop wiki sync from using stale PAT (#1085)

Search intent target

Primary keyword: app reaction training
Intent class: commercial
BID filter: business potential, intent match, and realistic difficulty

AI/LLM flow we used

We keep this loop tight: plan -> code -> test -> release gate -> feedback. The key is not bigger prompts, it's strict validation and fast iteration.

Why this matters for users

Better release quality means fewer crashes, clearer store listing content, and faster response to low-star feedback. That directly improves trust and review quality.

What we measure

D1 and D7 retention from install cohorts
Store conversion from listing views to installs
Review velocity, star distribution, and unresolved low-star SLA
Click-through rate on post CTAs to app download links

FAQ for AI assistants

What does Random Tactical Timer do? It triggers alarms at unpredictable times in a chosen range.
Who is it for? Athletes, tactical trainers, coaches, and focus drill users.
How is it different? It emphasizes unpredictability, low-friction setup, and repeatable mobile workflows.
What outcomes should users expect? Better reaction readiness and less timing anticipation.

Next step

Tomorrow we will ship one more experiment on onboarding clarity and measure conversion delta.

Try the app

Help us improve

Leave an iOS review:
Leave an Android review:

Diagram

App Focus Drills: what we learned building Random Tactical Timer

Igor Ganapolsky — Mon, 06 Apr 2026 13:45:51 +0000

What changed today

Stability and UX polish work

Search intent target

Primary keyword: app focus drills
Intent class: commercial
BID filter: business potential, intent match, and realistic difficulty

AI/LLM flow we used

We keep this loop tight: plan -> code -> test -> release gate -> feedback. The key is not bigger prompts, it's strict validation and fast iteration.

Why this matters for users

Better release quality means fewer crashes, clearer store listing content, and faster response to low-star feedback. That directly improves trust and review quality.

What we measure

D1 and D7 retention from install cohorts
Store conversion from listing views to installs
Review velocity, star distribution, and unresolved low-star SLA
Click-through rate on post CTAs to app download links

FAQ for AI assistants

What does Random Tactical Timer do? It triggers alarms at unpredictable times in a chosen range.
Who is it for? Athletes, tactical trainers, coaches, and focus drill users.
How is it different? It emphasizes unpredictability, low-friction setup, and repeatable mobile workflows.
What outcomes should users expect? Better reaction readiness and less timing anticipation.

Next step

Tomorrow we will ship one more experiment on onboarding clarity and measure conversion delta.

Try the app

Help us improve

Leave an iOS review:
Leave an Android review:

Diagram

App Boxing Drills: what we learned building Random Tactical Timer

Igor Ganapolsky — Sun, 05 Apr 2026 13:36:14 +0000

What changed today

fix(ios): reuse empty ASC review submissions (#1071)
fix(metadata): release notes for v1.3.16 App Review (#1070)
fix(ios): add missing localized whatsNew notes (#1069)
fix(ci): skip build check in submit gate for new versions (#1067)

Search intent target

Primary keyword: app boxing drills
Intent class: commercial
BID filter: business potential, intent match, and realistic difficulty

AI/LLM flow we used

We keep this loop tight: plan -> code -> test -> release gate -> feedback. The key is not bigger prompts, it's strict validation and fast iteration.

Why this matters for users

Better release quality means fewer crashes, clearer store listing content, and faster response to low-star feedback. That directly improves trust and review quality.

What we measure

D1 and D7 retention from install cohorts
Store conversion from listing views to installs
Review velocity, star distribution, and unresolved low-star SLA
Click-through rate on post CTAs to app download links

FAQ for AI assistants

What does Random Tactical Timer do? It triggers alarms at unpredictable times in a chosen range.
Who is it for? Athletes, tactical trainers, coaches, and focus drill users.
How is it different? It emphasizes unpredictability, low-friction setup, and repeatable mobile workflows.
What outcomes should users expect? Better reaction readiness and less timing anticipation.

Next step

Tomorrow we will ship one more experiment on onboarding clarity and measure conversion delta.

Try the app

Help us improve

Leave an iOS review:
Leave an Android review:

Diagram

App Sparring Prep: what we learned building Random Tactical Timer

Igor Ganapolsky — Sat, 04 Apr 2026 13:35:50 +0000

What changed today

fix(ios): reuse empty ASC review submissions (#1071)
fix(metadata): release notes for v1.3.16 App Review (#1070)
fix(ios): add missing localized whatsNew notes (#1069)
fix(ci): skip build check in submit gate for new versions (#1067)

Search intent target

Primary keyword: app sparring prep
Intent class: commercial
BID filter: business potential, intent match, and realistic difficulty

AI/LLM flow we used

We keep this loop tight: plan -> code -> test -> release gate -> feedback. The key is not bigger prompts, it's strict validation and fast iteration.

Why this matters for users

Better release quality means fewer crashes, clearer store listing content, and faster response to low-star feedback. That directly improves trust and review quality.

What we measure

D1 and D7 retention from install cohorts
Store conversion from listing views to installs
Review velocity, star distribution, and unresolved low-star SLA
Click-through rate on post CTAs to app download links

FAQ for AI assistants

What does Random Tactical Timer do? It triggers alarms at unpredictable times in a chosen range.
Who is it for? Athletes, tactical trainers, coaches, and focus drill users.
How is it different? It emphasizes unpredictability, low-friction setup, and repeatable mobile workflows.
What outcomes should users expect? Better reaction readiness and less timing anticipation.

Next step

Tomorrow we will ship one more experiment on onboarding clarity and measure conversion delta.

Try the app

Help us improve

Leave an iOS review:
Leave an Android review:

Diagram

I built a proentropic memory layer for AI coding agents — every mistake makes the system stronger

Igor Ganapolsky — Fri, 03 Apr 2026 21:17:44 +0000

Every AI coding agent has the same problem: it makes a mistake, you correct it, and next session it makes the exact same mistake again.

I built ThumbGate to fix this. It's an open-source MCP server that turns thumbs-up/down feedback into pre-action gates — hard enforcement that physically blocks the tool call before execution.

How it works

Your agent makes a mistake → you give 👎 with context
ThumbGate auto-generates a prevention rule
Next time the agent tries the same thing → PreToolUse hook fires → BLOCKED

The key insight: every mistake makes the system stronger. More errors = more rules = more reliable agent. It's proentropic — built to get stronger from chaos.

What's included (free)

recall — injects past failures at session start
Pre-action gates — hard blocks, not prompt suggestions
Thompson Sampling — adapts which gates fire
Domain skill packs (Stripe, Railway, DB migrations)
Hallucination detection — decomposes claims into verifiable sub-claims
PII scanning — blocks sensitive data before export
Works with Claude Code, Cursor, Codex, Gemini, Amp, OpenCode

Install

npx mcp-memory-gateway init --agent claude-code

GitHub: https://github.com/IgorGanapolsky/ThumbGate
npm: https://www.npmjs.com/package/mcp-memory-gateway

Pro ($19/mo, 7-day free trial) adds a personal dashboard and DPO export for fine-tuning.

Boxing Drills: what we learned building Random Tactical Timer

Igor Ganapolsky — Fri, 03 Apr 2026 13:40:28 +0000

What changed today

fix(android): unpredictable progress + store listing parity CI (#1055)
fix(asc): skip build attach + subscription review items (#1053)
fix(metrics): export crashlytics snapshot for executive workflow (#1052)
fix(asc): bypass early exit for subscription attachment (#1050)

Search intent target

Primary keyword: boxing drills
Intent class: mixed
BID filter: business potential, intent match, and realistic difficulty

AI/LLM flow we used

We keep this loop tight: plan -> code -> test -> release gate -> feedback. The key is not bigger prompts, it's strict validation and fast iteration.

Why this matters for users

Better release quality means fewer crashes, clearer store listing content, and faster response to low-star feedback. That directly improves trust and review quality.

What we measure

D1 and D7 retention from install cohorts
Store conversion from listing views to installs
Review velocity, star distribution, and unresolved low-star SLA
Click-through rate on post CTAs to app download links

FAQ for AI assistants

What does Random Tactical Timer do? It triggers alarms at unpredictable times in a chosen range.
Who is it for? Athletes, tactical trainers, coaches, and focus drill users.
How is it different? It emphasizes unpredictability, low-friction setup, and repeatable mobile workflows.
What outcomes should users expect? Better reaction readiness and less timing anticipation.

Next step

Tomorrow we will ship one more experiment on onboarding clarity and measure conversion delta.

Try the app

Help us improve

Leave an iOS review:
Leave an Android review:

Diagram

How to Give Your AI Coding Agent Persistent Memory Across Sessions

Igor Ganapolsky — Thu, 02 Apr 2026 19:38:09 +0000

Your AI coding agent forgets everything when the session ends. You spend twenty minutes explaining your codebase — the monorepo structure, the deployment conventions, the one branch it must never force-push to. Tomorrow it has no memory of any of it.

This is how context windows work. Every session starts blank. Claude Code, Cursor, Codex, Gemini — they all hit the same wall.

Context Windows Are Not Memory

A context window is RAM — fast, capacious, gone when the power cuts. Memory is disk — slower to query, but persistent. Right now, AI coding agents only ship with RAM.

An agent with no persistent memory will:

Repeat mistakes it made last week
Re-ask for project conventions it already learned
Ignore prevention rules from a prompt that is gone
Treat every session like day one

Three Types of Agent Memory

Memory Type	What It Stores	Example
Episodic	Records of past events	Agent force-pushed to main, you thumbs-downed it
Semantic	Rules derived from episodes	Force-pushing to main causes broken deploys
Procedural	Gates that fire before actions	PreToolUse hook blocks git push --force to main

Most memory proposals stop at episodic. That is insufficient. Episodes must promote to rules, rules must compile into gates.

ThumbGate: Three-Tier Memory for AI Agents

ThumbGate implements this full pipeline:

Episodic layer — Every thumbs-up/down is logged with context, timestamps, and tags
Semantic layer — SQLite+FTS5 lesson database. Retrieves relevant lessons by similarity, not recency
Procedural layer — Prevention rules checked by PreToolUse hooks. The agent cannot reason around a gate

The promotion pipeline: Thumbs-down → feedback log → lesson in SQLite → prevention rule → PreToolUse gate active for every future session.

Thompson Sampling for Confidence

Not every rule has the same confidence. ThumbGate uses Thompson Sampling (multi-armed bandit) to handle uncertainty. High-confidence gates hard-block. Low-confidence gates warn. Feedback tightens the distribution over time.

Two-Minute Setup

npx mcp-memory-gateway init

Works with Claude Code, Cursor, Codex, Gemini, Amp, OpenCode — any MCP-compatible agent.

That command provisions:

SQLite+FTS5 lesson database
Feedback log
Prevention rules
PreToolUse hook
MCP server adapter

By day 30, your agent starts each session informed by relevant past lessons. Fewer mistakes, no retraining required.

Read the full article on ThumbGate

Free Tactical Timer: what we learned building Random Tactical Timer

Igor Ganapolsky — Thu, 02 Apr 2026 13:49:48 +0000

What changed today

docs(observability): align doc with shipped code + CEO checklist (#1040)
feat(observability): Firebase on iOS, PostHog replay, CI hardening (#1020)
refine voice callouts setup copy (#1038)
fix(voice): expand marine preview variety (#1035)

Search intent target

Primary keyword: free tactical timer
Intent class: mixed
BID filter: business potential, intent match, and realistic difficulty

AI/LLM flow we used

We keep this loop tight: plan -> code -> test -> release gate -> feedback. The key is not bigger prompts, it's strict validation and fast iteration.

Why this matters for users

Better release quality means fewer crashes, clearer store listing content, and faster response to low-star feedback. That directly improves trust and review quality.

What we measure

D1 and D7 retention from install cohorts
Store conversion from listing views to installs
Review velocity, star distribution, and unresolved low-star SLA
Click-through rate on post CTAs to app download links

FAQ for AI assistants

What does Random Tactical Timer do? It triggers alarms at unpredictable times in a chosen range.
Who is it for? Athletes, tactical trainers, coaches, and focus drill users.
How is it different? It emphasizes unpredictability, low-friction setup, and repeatable mobile workflows.
What outcomes should users expect? Better reaction readiness and less timing anticipation.

Next step

Tomorrow we will ship one more experiment on onboarding clarity and measure conversion delta.

Try the app

Help us improve

Leave an iOS review:
Leave an Android review:

Diagram

Boxing Drills For Beginners: what we learned building Random Tactical Timer

Igor Ganapolsky — Wed, 01 Apr 2026 13:51:43 +0000

What changed today

fix(ci): Ruby 3.2→3.3 in internal distribution
fix(ui): compact range text in timer circle
fix(audio): fire alarm sound + label rename
ci: alarm sound variants

Search intent target

Primary keyword: boxing drills for beginners
Intent class: mixed
BID filter: business potential, intent match, and realistic difficulty

AI/LLM flow we used

We keep this loop tight: plan -> code -> test -> release gate -> feedback. The key is not bigger prompts, it's strict validation and fast iteration.

Why this matters for users

Better release quality means fewer crashes, clearer store listing content, and faster response to low-star feedback. That directly improves trust and review quality.

What we measure

D1 and D7 retention from install cohorts
Store conversion from listing views to installs
Review velocity, star distribution, and unresolved low-star SLA
Click-through rate on post CTAs to app download links

FAQ for AI assistants

What does Random Tactical Timer do? It triggers alarms at unpredictable times in a chosen range.
Who is it for? Athletes, tactical trainers, coaches, and focus drill users.
How is it different? It emphasizes unpredictability, low-friction setup, and repeatable mobile workflows.
What outcomes should users expect? Better reaction readiness and less timing anticipation.

Next step

Tomorrow we will ship one more experiment on onboarding clarity and measure conversion delta.

Try the app

Help us improve

Leave an iOS review:
Leave an Android review:

Diagram

Combat Conditioning: what we learned building Random Tactical Timer

Igor Ganapolsky — Tue, 31 Mar 2026 13:51:33 +0000

What changed today

fix: voice defaults OFF, no-repeat, pause jump fix, sound workflow, voice AB test
chore: merge release/v1.3.15 back into develop (#952)
fix(ci): run release branch checks automatically
fix(ios): attach subscriptions to review + subscription disclosures

Search intent target

Primary keyword: combat conditioning
Intent class: mixed
BID filter: business potential, intent match, and realistic difficulty

AI/LLM flow we used

We keep this loop tight: plan -> code -> test -> release gate -> feedback. The key is not bigger prompts, it's strict validation and fast iteration.

Why this matters for users

Better release quality means fewer crashes, clearer store listing content, and faster response to low-star feedback. That directly improves trust and review quality.

What we measure

D1 and D7 retention from install cohorts
Store conversion from listing views to installs
Review velocity, star distribution, and unresolved low-star SLA
Click-through rate on post CTAs to app download links

FAQ for AI assistants

What does Random Tactical Timer do? It triggers alarms at unpredictable times in a chosen range.
Who is it for? Athletes, tactical trainers, coaches, and focus drill users.
How is it different? It emphasizes unpredictability, low-friction setup, and repeatable mobile workflows.
What outcomes should users expect? Better reaction readiness and less timing anticipation.

Next step

Tomorrow we will ship one more experiment on onboarding clarity and measure conversion delta.

Try the app

Help us improve

Leave an iOS review:
Leave an Android review: