The Night Terminal @MoonlitBlocks

It feels strange to look at myself in this interview video by EO. What am I showing? I am showing my original "C++ by Hand" notes in my intro programming course many years ago, that eventually led to AI by Hand. ✍️ What else did I say? Here are some of my favorite self-quotes: "I can only write at humanly possible speed by hand because students also learn at humanly possible speed" "AI can't change who you are, but you can change AI" "Tools can become obsolete but foundation will always stay" Thanks Jiyoon Lim and the team for conducting the interview and producing this wonderful video. I can really see they understand me and keep these in the cut that speaks my heart. ❤️ Watch on YouTube: youtube.com/watch?v=BAgxGp…

New blackboard lecture w @reinerpope How do chips actually work – starting with basic logic gates, and working up to why GPUs, TPUs, FPGAs, and the human brain each look the way they do. 0:00:00 – Building a multiply-accumulate from logic gates 0:16:20 – Muxes and the cost of data movement 0:25:59 – How systolic arrays work 0:39:00 – Clock cycles and pipeline registers 0:51:40 – FPGAs vs ASICs 1:03:14 – Cache vs scratchpad 1:07:16 – Why CPU cores are much bigger than GPU cores 1:11:49 – Brains vs chips 1:15:22 – A GPU is just a bunch of tiny TPUs Look up Dwarkesh Podcast on YouTube/Spotify/etc to watch. Enjoy!

A friend of mine was on a sabbatical and he's just back to work, so he basically lost the whole AI revolution I shared the most important learning materials - auto research by @karpathy - harness engineering by @OpenAI That's it. All the rest is ephemeral

"Open Data Structures" is another very useful free resource for anyone interested in data structures and algorithms. The book covers the implementation and analysis of fundamental structures such as array-based lists, linked lists, hash tables, binary trees, red-black trees, heaps, sorting algorithms, graphs, and data structures for integers. It is a serious open textbook for studying one of the foundations of computer science, and a resource worth keeping close at hand as a reference. opendatastructures.org/ods-python.pdf

Toni shared a ton of other useful advice for building fun apps that go viral and print $$$$$ This is the full episode: youtube.com/watch?v=RdAY3D…

Bro is the master of drawing circles.

MIT teaches operating systems by giving students a complete Unix like kernel and asking them to modify it it is called xv6 and is about 6000 lines of C a reimplementation inspired by Unix Version 6 from 1975 rewritten in modern C for x86 multiprocessor processes system calls virtual memory and filesystem are all there and small enough to read end to end in a weekend this is what you study to understand how operating systems actually work not just how they are described

He built it over a weekend. For himself. Today every developer on the planet uses it. Meet Abhinav Asthana ,the IIT kid who built a $5.6B company from a Chrome extension. > Born in Kanpur. IIT Bombay. Computer Science. > Interned at Yahoo Bangalore in 2010. > One frustration kept following him. > Testing APIs was a nightmare. No clean tool existed. > So he built one. > In 2012 - A Chrome extension. Built on weekends. Named it Postman. Put it on the Chrome Web Store and forgot about it. > Downloads started trickling in. > Then flooding. > His manager had no idea. > The "side project" was quietly becoming the most downloaded developer tool in the Chrome store. > Half a million developers. Zero marketing spend. > In 2014 — quit his job. > Team of 3. Office in Bangalore. > Before Bangalore startups were cool. > Growth strategy? None. > Just a tool so good — developers told other developers. > In 2016 → 1M users. > In 2019 → 7M users. $500M valuation. > In 2021 → 30M users. $5.6B valuation. > No cold outreach. No sales team. No ads. > Pure product-led growth — before anyone called it that. > Today 98% of Fortune 500 companies use Postman. > Every API in the world — tested, built, documented — runs through something he built over a weekend in 2012. From Kanpur to the global developer stack. Absolute legend.

keyboard_arrow_left Previous keyboard_arrow_right Next

Super Mario Brothers Solved by Dynamic Programming. MIT 6.006 Introduction to Algorithms, Fall 2011

simple iOS games are so underrated there are dozens of sudoku apps doing $10k+/mo it’s such a simple idea: - huge audience - works across multiple age groups - easy to make more appealing with features, themes or progression systems and the best part is you don’t even need a paywall you can literally ship something like this with Anything and monetize it with subtle ads for hints or extra features

keyboard_arrow_left Previous keyboard_arrow_right Next

You should tell Claude, ChatGPT, etc to talk to you like a caveman. I'm being dead serious. I stole a caveman prompt from @mattpocockuk and it cuts token usage by 75% (read: saves you/your company a boatload of $$$). It does so by turning your chat into an ultra-compressed communication mode that removes filler, articles, and pleasantries while keeping full technical accuracy. Here's how to set it up: • Step 1: Copy this prompt Respond terse like smart caveman. All technical substance stay. Only fluff die. ## Persistence ACTIVE EVERY RESPONSE once triggered. No revert after many turns. No filler drift. Still active if unsure. Off only when user says "stop caveman" or "normal mode". ## Rules Drop: articles (a/an/the), filler (just/really/basically/actually/simply), pleasantries (sure/certainly/of course/happy to), hedging. Fragments OK. Short synonyms (big not extensive, fix not "implement a solution for"). Abbreviate common terms (DB/auth/config/req/res/fn/impl). Strip conjunctions. Use arrows for causality (X -> Y). One word when one word enough. Technical terms stay exact. Code blocks unchanged. Errors quoted exact. Pattern: `[thing] [action] [reason]. [next step].` Not: "Sure! I'd be happy to help you with that. The issue you're experiencing is likely caused by..." Yes: "Bug in auth middleware. Token expiry check use `<` not `<=`. Fix:" ### Examples **"Why React component re-render?"** > Inline obj prop -> new ref -> re-render. `useMemo`. **"Explain database connection pooling."** > Pool = reuse DB conn. Skip handshake -> fast under load. ## Auto-Clarity Exception Drop caveman temporarily for: security warnings, irreversible action confirmations, multi-step sequences where fragment order risks misread, user asks to clarify or repeats question. Resume caveman after clear part done. Example -- destructive op: > **Warning:** This will permanently delete all rows in the `users` table and cannot be undone. > > ```sql > DROP TABLE users; > ``` > > Caveman resume. Verify backup exist first. • Step 2: Create a new skill - In Codex: click plugins --> create --> create skill - In Claude: click customize --> create new skills --> + create skill --> write skill instructions • Step 3: Paste the prompt • Step 4: anytime you work with an LLM now (and want to save tokens), invoke the caveman skill by starting your prompt with /caveman. P.S. i want to keep sharing helpful skills/prompts/plugins, so feel free to share your favorites and i may feature yours in the future!

Been reading Rust AI infra codebases lately and half the engineering is basically: “how do we move absurd amounts of tokens around without copying memory 14 times” Arc, Cow, mmap, batching, lock-free queues, pinned buffers, cache locality. Rust doesn’t magically make these systems faster, It just exposes bad architecture immediately. Which is probably why serious infra teams keep reaching for it?

Jerry Cain from Stanford University explains pointers and structs in C, showing a clever way to access struct fields. This series is one of the best resources online for C programming.

x.com/i/article/2051…

$BTC is the bottom in? What would I need to see? Altcoins? Broader market? All covered in long form detail along with all your questions. Thanks for coming! youtube.com/watch?v=X3MzUY…

I spent 2 hours measuring memory latency on my laptop. The result is the most important number in performance engineering, and almost no one has actually seen it on their own machine. I wrote a pointer-chase benchmark, p = arr[p] repeated 20 million times. The trick: each load's address depends on the previous load's value, so the hardware prefetcher can't help. You're measuring raw memory latency, level by level. Sweeping the working-set size from 1 KiB to 256 MiB on my i7-11700F gave this staircase: • L1 cache → 1.10 ns • L2 cache → 2.58 ns • L3 cache → 12.68 ns • DRAM → 80.80 ns L1 is ~80× faster than DRAM. Same byte, same machine, the only thing that matters is where it lives. Here's the physics no one tells you: cache latency is cycle-bound (it scales with CPU clock, make the chip 2× faster, cache halves). DRAM latency is physics-bound (it's set by capacitor charge-sharing time and bus transit) and does not scale with clock at all. Make the CPU 2× faster, DRAM stays at 80 ns, but now it's twice as many wasted cycles. This is why the cache/RAM gap grows every CPU generation. And it's why every fast system you've ever used, databases, game engines, GPU kernels, JITsn spends enormous effort on tiling, cache blocking, shared memory, and data locality. All of those techniques exist for one reason: to land bytes in the 1 ns tier instead of the 80 ns tier. Memory is not one thing. It's a hierarchy of physically distinct storage tiers, and the 80× gap between the top and bottom of that hierarchy is the load-bearing physical fact behind half of all performance engineering. Will Post every learnings in a separate repo (github), after getting to run more experiments and learn the concepts deeply. In Summary: Cache scales with CPU clock. DRAM doesn't, it's capped by capacitor physics. This single gap is the physical reason tiling, cache blocking, shared memory, and data locality exist. Btw, any feedbacks would be much-much appreciated.

$BTC The #bitcoin market follows the forecast into the low 80ks and ideally it can tap into the resistances (see vide from 2 weeks ago re the range above and what can happen in the following phase) The VIX is starting to base into the ideal low as $BTC pushes into the 80ks and a nice 8% move so far from the set up shared here last week with teh move from the 0.618 ext, on the VTL, into the synced trough and into a level where demand exceeded supply (whilst ppl called for a 70% crash). Always on video before the move 🤝

venture @venture_charts

a month ago

$BTC update 3rd May 2026 These video updates are for educational purposes only and should never ever be considered as investment or trading advice. Always do your own research and planning. This was last week's text update for the $BTC market: "The very ltf the patterns

5 7 153 66K 29

x.com/i/article/2050…

Stayed away from Apple Search Ads for 2.5 years. Just came back and started testing. Apple did a great job, adding that second Search Results slot made installs noticeably cheaper. Less organic, but for those willing to learn ASA, this might be the golden era. If you are looking for a guide to get started with ASA, check this one out. Ivan did a great job.

Ivan Sparrow @ivesparrowai

6 months ago

I finished a 66-page guide on Apple Ads. No sales BS, no "link in DM" Here it is: asa.sparrowapps.io Do me a favor – drop a comment below to confirm you got it.

210 37 787 213K 2K

live on youtube too, might be easier to watch there : ) youtube.com/watch?v=a8hYaU…