Jakub Slys
@iam.slys.dev
𝑇ℎ𝑒 𝑏𝑒𝑠𝑡 𝑡𝑖𝑚𝑒 𝑡𝑜 𝑝𝑙𝑎𝑛𝑡 𝑎 𝑡𝑟𝑒𝑒 𝑤𝑎𝑠 20 𝑦𝑒𝑎𝑟𝑠 𝑎𝑔𝑜. 𝑻𝒉𝒆 𝒔𝒆𝒄𝒐𝒏𝒅 𝒃𝒆𝒔𝒕 𝒕𝒊𝒎𝒆 𝒊𝒔 𝒏𝒐𝒘.
Doing more isn’t always the goal—doing what matters is. When automation clears the noise, your ideas finally get the room to breathe. If you could automate one creative hassle, what would it be?
How you can automate sharing your Substack to X, Threads, Notes & Bluesky
A journey into protecting your creative spark - by letting a system carry the load
iam.slys.dev
January 11, 2026 at 3:16 PM
Doing more isn’t always the goal—doing what matters is. When automation clears the noise, your ideas finally get the room to breathe. If you could automate one creative hassle, what would it be?
YTPro’s tiny APK still delivers captions, downloads, gestures, and even Gemini-powered summaries. It’s fun to see a project push both technology and nostalgia at once. What features would you build for legacy users?
devtools retro ai productivity
devtools retro ai productivity
GitHub - prateek-chaubey/YTPro: Youtube client with older Android version support, background player, Google Gemini ✨ and many more features.
Youtube client with older Android version support, background player, Google Gemini ✨ and many more features. - prateek-chaubey/YTPro
github.com
January 11, 2026 at 12:14 PM
YTPro’s tiny APK still delivers captions, downloads, gestures, and even Gemini-powered summaries. It’s fun to see a project push both technology and nostalgia at once. What features would you build for legacy users?
devtools retro ai productivity
devtools retro ai productivity
Reverse engineering is about curiosity, respect, and making things better for everyone. Tinker ethically, automate what inspires you, and share your story. What should be my next target? Read more: https://iam.slys.dev/p/no-official-api-no-problem-how-i
No official API⁉️ No problem‼️ How I reverse-engineered Substack API (and so can you)
You don’t need official docs to create real tools - just curiosity and patience
iam.slys.dev
January 10, 2026 at 3:16 PM
Reverse engineering is about curiosity, respect, and making things better for everyone. Tinker ethically, automate what inspires you, and share your story. What should be my next target? Read more: https://iam.slys.dev/p/no-official-api-no-problem-how-i
Ever wanted a search agent that can do more than just look things up? MiroThinker’s interactive scaling gives it a wild ability: hundreds of tool uses, big context, and deep reasoning. Open-source research just got a major upgrade. Imagine what projects you’d build with it!
opensource AI research
opensource AI research
GitHub - MiroMindAI/MiroThinker: MiroThinker is an open-source search agent model, built for tool-augmented reasoning and real-world information seeking, aiming to match the deep research experience of OpenAI Deep Research and Gemini Deep Research.
MiroThinker is an open-source search agent model, built for tool-augmented reasoning and real-world information seeking, aiming to match the deep research experience of OpenAI Deep Research and Gem...
github.com
January 10, 2026 at 12:14 PM
Ever wanted a search agent that can do more than just look things up? MiroThinker’s interactive scaling gives it a wild ability: hundreds of tool uses, big context, and deep reasoning. Open-source research just got a major upgrade. Imagine what projects you’d build with it!
opensource AI research
opensource AI research
Ever noticed how some apps feel smooth, even if the data isn’t current everywhere? That’s optimistic UI helping users feel the system is reliable, masking any sync lag behind thoughtful design. Where have you seen this work well?
What it means for a system to be consistent — and why it does not always have to be
System design fundamentals
iam.slys.dev
January 9, 2026 at 8:21 PM
Ever noticed how some apps feel smooth, even if the data isn’t current everywhere? That’s optimistic UI helping users feel the system is reliable, masking any sync lag behind thoughtful design. Where have you seen this work well?
Early estimation isn’t guesswork—it’s how you dodge disasters. Do you run napkin math before building, or only after something breaks at scale?
From TinyURL to Bitly (and beyond): designing a smarter URL shortener
Principles and practice of system design
iam.slys.dev
January 9, 2026 at 3:16 PM
Early estimation isn’t guesswork—it’s how you dodge disasters. Do you run napkin math before building, or only after something breaks at scale?
Trying out MiroThinker feels like the future: transparent agent configs, code execution, and real-time web search. Makes me wonder—how will workflows shift when everyone gets a smart AI research assistant for free?
LLM agents ML
LLM agents ML
GitHub - MiroMindAI/MiroThinker: MiroThinker is an open-source search agent model, built for tool-augmented reasoning and real-world information seeking, aiming to match the deep research experience of OpenAI Deep Research and Gemini Deep Research.
MiroThinker is an open-source search agent model, built for tool-augmented reasoning and real-world information seeking, aiming to match the deep research experience of OpenAI Deep Research and Gem...
github.com
January 9, 2026 at 12:14 PM
Trying out MiroThinker feels like the future: transparent agent configs, code execution, and real-time web search. Makes me wonder—how will workflows shift when everyone gets a smart AI research assistant for free?
LLM agents ML
LLM agents ML
Splitting means query complexity, extra maintenance, and the never-ending chore of balancing hot spots. Where does your stack groan under growth? How do you handle the trade-offs? Read more: https://iam.slys.dev/p/why-databases-get-split-sharding
Why databases get split — sharding, partitioning, and replication without fear
System design fundamentals
iam.slys.dev
January 8, 2026 at 8:21 PM
Splitting means query complexity, extra maintenance, and the never-ending chore of balancing hot spots. Where does your stack groan under growth? How do you handle the trade-offs? Read more: https://iam.slys.dev/p/why-databases-get-split-sharding
Frozen time-series foundation models are starting to appear as toolkits—suddenly, classification and anomaly detection tasks could be plug-and-play. Would you shift away from bespoke pipelines if these work? Why or why not?
✨ Scientific ML leads, time-series foundation models generalize, and real-world agent benchmarks temper automation hype
AI Alert
iam.slys.dev
January 8, 2026 at 3:16 PM
Frozen time-series foundation models are starting to appear as toolkits—suddenly, classification and anomaly detection tasks could be plug-and-play. Would you shift away from bespoke pipelines if these work? Why or why not?
Replacing heavy RAG pipelines with a serverless, model-agnostic memory file feels like a new chapter for AI agents. Memvid’s Smart Frames and time-travel debugging could open up easier prototyping and more trustworthy AI.
serverless opensource agents
serverless opensource agents
GitHub - memvid/memvid: Memory layer for AI Agents. Replace complex RAG pipelines with a serverless, single-file memory layer. Give your agents instant retrieval and long-term memory.
Memory layer for AI Agents. Replace complex RAG pipelines with a serverless, single-file memory layer. Give your agents instant retrieval and long-term memory. - memvid/memvid
github.com
January 8, 2026 at 12:14 PM
Replacing heavy RAG pipelines with a serverless, model-agnostic memory file feels like a new chapter for AI agents. Memvid’s Smart Frames and time-travel debugging could open up easier prototyping and more trustworthy AI.
serverless opensource agents
serverless opensource agents
L1/L2/L3—arcane labels, real impact. Each cache level is a trade-off, making your apps fast if data stays close. Cache misses mean your CPU queues up for slow RAM. How do you design for cache friendliness? More thoughts: https://iam.slys.dev/p/cpu-caches-why-tiny-memory-matters
CPU Caches: Why tiny memory matters?
System design
iam.slys.dev
January 7, 2026 at 3:16 PM
L1/L2/L3—arcane labels, real impact. Each cache level is a trade-off, making your apps fast if data stays close. Cache misses mean your CPU queues up for slow RAM. How do you design for cache friendliness? More thoughts: https://iam.slys.dev/p/cpu-caches-why-tiny-memory-matters
Cloud gaming on your own hardware? Sunshine makes it possible. Less lag, no subscription, and open-source foundation—what’s not to love? Have you tried setting up your own game streamer yet?
selfhosted techdiy gamestreaming freedom
selfhosted techdiy gamestreaming freedom
GitHub - LizardByte/Sunshine: Self-hosted game stream host for Moonlight.
Self-hosted game stream host for Moonlight. Contribute to LizardByte/Sunshine development by creating an account on GitHub.
github.com
January 7, 2026 at 12:14 PM
Cloud gaming on your own hardware? Sunshine makes it possible. Less lag, no subscription, and open-source foundation—what’s not to love? Have you tried setting up your own game streamer yet?
selfhosted techdiy gamestreaming freedom
selfhosted techdiy gamestreaming freedom
Less guessing, more knowing: MCP lets models access live data and external tools, making answers sharper and more reliable. What’s something new you’d trust your AI to do now? Read more: https://iam.slys.dev/p/understanding-the-model-context-protocol
Understanding the Model Context Protocol in Artificial Intelligence
Machine Learning and Artificial Intelligence
iam.slys.dev
January 6, 2026 at 8:21 PM
Less guessing, more knowing: MCP lets models access live data and external tools, making answers sharper and more reliable. What’s something new you’d trust your AI to do now? Read more: https://iam.slys.dev/p/understanding-the-model-context-protocol
It often takes a tiny change—like reaching for a hash map—to make code dramatically better. Have you ever explained this approach to someone new to coding?
🧠 Crack Google Interview: The sum of two
Google LeetCode
iam.slys.dev
January 6, 2026 at 6:06 PM
It often takes a tiny change—like reaching for a hash map—to make code dramatically better. Have you ever explained this approach to someone new to coding?
Edge DNS isn’t just theory—serverless-dns makes privacy-first, content-blocking resolvers deployable anywhere in the cloud. 190+ blocklists and less than 2ms server latency… Do you trust the internet more when you own your infra?
serverless privacy DNS cloud
serverless privacy DNS cloud
GitHub - serverless-dns/serverless-dns: The RethinkDNS resolver that deploys to Cloudflare Workers, Deno Deploy, Fastly, and Fly.io
The RethinkDNS resolver that deploys to Cloudflare Workers, Deno Deploy, Fastly, and Fly.io - serverless-dns/serverless-dns
github.com
January 2, 2026 at 12:13 PM
Edge DNS isn’t just theory—serverless-dns makes privacy-first, content-blocking resolvers deployable anywhere in the cloud. 190+ blocklists and less than 2ms server latency… Do you trust the internet more when you own your infra?
serverless privacy DNS cloud
serverless privacy DNS cloud
Knowledge quite literally pays off. I’m launching “Get Your Cert”—refer friends, and you could get help funding your next cert. Let’s grow our community and our careers together! Curious about the details? Read more: https://iam.slys.dev/p/get-your-cert
Get Your Cert
💰 Knowledge pays 💰 Literally.
iam.slys.dev
January 1, 2026 at 8:22 PM
Knowledge quite literally pays off. I’m launching “Get Your Cert”—refer friends, and you could get help funding your next cert. Let’s grow our community and our careers together! Curious about the details? Read more: https://iam.slys.dev/p/get-your-cert
The median problem is really about divide-and-conquer. Instead of sorting everything, it's about finding that clean split between left and right. Have you had an "aha" moment rethinking where to cut?
Read more: https://iam.slys.dev/p/crack-google-interview-can-you-finding
Read more: https://iam.slys.dev/p/crack-google-interview-can-you-finding
🧠 Crack Google Interview: Can you finding the median of two sorted arrays?
Google LeetCode
iam.slys.dev
December 31, 2025 at 8:22 PM
The median problem is really about divide-and-conquer. Instead of sorting everything, it's about finding that clean split between left and right. Have you had an "aha" moment rethinking where to cut?
Read more: https://iam.slys.dev/p/crack-google-interview-can-you-finding
Read more: https://iam.slys.dev/p/crack-google-interview-can-you-finding
Ever wish you never had to do the email-code shuffle during signups? The Email Verification Protocol is aiming for just that—fast, private, and seamless identity proofs powered by your browser and email domain. Game changer?
privacy onboarding tech
privacy onboarding tech
GitHub - WICG/email-verification-protocol: verified autofill
verified autofill . Contribute to WICG/email-verification-protocol development by creating an account on GitHub.
github.com
December 31, 2025 at 12:13 PM
Ever wish you never had to do the email-code shuffle during signups? The Email Verification Protocol is aiming for just that—fast, private, and seamless identity proofs powered by your browser and email domain. Game changer?
privacy onboarding tech
privacy onboarding tech
If your tech stack juggles platforms and apps, SOA helps orchestrate complexity. But is it too much for simple projects? When has SOA been overkill for you?
Service-Oriented Architecture
Software architecture
iam.slys.dev
December 30, 2025 at 3:17 PM
If your tech stack juggles platforms and apps, SOA helps orchestrate complexity. But is it too much for simple projects? When has SOA been overkill for you?
What impressed me about LightRAG: support for multimodal data, knowledge graph editing, plug-n-play LLMs, and detailed docs. Feels like a playground for retrieval-augmented innovation. Anyone else tried mixing custom storage backends in production?
opensource genai vectorsearch llm
opensource genai vectorsearch llm
GitHub - HKUDS/LightRAG: [EMNLP2025] "LightRAG: Simple and Fast Retrieval-Augmented Generation"
[EMNLP2025] "LightRAG: Simple and Fast Retrieval-Augmented Generation" - HKUDS/LightRAG
github.com
December 30, 2025 at 12:13 PM
What impressed me about LightRAG: support for multimodal data, knowledge graph editing, plug-n-play LLMs, and detailed docs. Feels like a playground for retrieval-augmented innovation. Anyone else tried mixing custom storage backends in production?
opensource genai vectorsearch llm
opensource genai vectorsearch llm
Scaling in SOA means picking what matters most. Got a traffic spike? Scale just that service! It’s all about smart modularity. Where could targeted scaling have spared you a headache?
Service-Oriented Architecture
Software architecture
iam.slys.dev
December 29, 2025 at 8:22 PM
Scaling in SOA means picking what matters most. Got a traffic spike? Scale just that service! It’s all about smart modularity. Where could targeted scaling have spared you a headache?
Why use TypeScript? Because strong types make even mystery data feel reliable. With my Substack API client, what starts as guesswork becomes safe, functional code. Do types help you think more clearly? Read more: https://iam.slys.dev/p/no-official-api-no-problem-how-i
No official API⁉️ No problem‼️ How I reverse-engineered Substack API (and so can you)
You don’t need official docs to create real tools - just curiosity and patience
iam.slys.dev
December 29, 2025 at 3:17 PM
Why use TypeScript? Because strong types make even mystery data feel reliable. With my Substack API client, what starts as guesswork becomes safe, functional code. Do types help you think more clearly? Read more: https://iam.slys.dev/p/no-official-api-no-problem-how-i
Modern web users juggle a lot—so why not let a tool handle the nagging task of page monitoring? changedetection.io alerts you to changes, big or small. Imagine getting notified the moment it counts! How would you use this freedom?
infosec monitoring automation
infosec monitoring automation
GitHub - dgtlmoon/changedetection.io: Best and simplest tool for website change detection, web page monitoring, and website change alerts. Perfect for tracking content changes, price drops, restock alerts, and website defacement monitoring—all for free or enjoy our SaaS plan!
Best and simplest tool for website change detection, web page monitoring, and website change alerts. Perfect for tracking content changes, price drops, restock alerts, and website defacement monito...
github.com
December 29, 2025 at 12:13 PM
Modern web users juggle a lot—so why not let a tool handle the nagging task of page monitoring? changedetection.io alerts you to changes, big or small. Imagine getting notified the moment it counts! How would you use this freedom?
infosec monitoring automation
infosec monitoring automation
Ever notice how long a thread holds a lock? The longer the hold, the bigger the queue behind it. I've been reflecting on how making critical sections small can free up everyone. Have you tuned a critical section lately? Read more: https://iam.slys.dev/p/understanding-locking-contention
Understanding locking contention in computing
System Design
iam.slys.dev
December 28, 2025 at 3:17 PM
Ever notice how long a thread holds a lock? The longer the hold, the bigger the queue behind it. I've been reflecting on how making critical sections small can free up everyone. Have you tuned a critical section lately? Read more: https://iam.slys.dev/p/understanding-locking-contention
I’m fascinated by the idea of browsers verifying email ownership without ever leaving the page or leaking what apps you use. This protocol’s use of cryptographic tokens plus DNS and browser APIs feels like pure UX evolution. How might this reshape sign-ups?
security UX browser
security UX browser
GitHub - WICG/email-verification-protocol: verified autofill
verified autofill . Contribute to WICG/email-verification-protocol development by creating an account on GitHub.
github.com
December 28, 2025 at 12:13 PM
I’m fascinated by the idea of browsers verifying email ownership without ever leaving the page or leaking what apps you use. This protocol’s use of cryptographic tokens plus DNS and browser APIs feels like pure UX evolution. How might this reshape sign-ups?
security UX browser
security UX browser