Mark Papadakis

Seeking Knowledge 24x7. Software Engineer — Family, Technology, Books, Basketball, Video Games.

[Mark Papadakis @markpapadakis]

@TheNatural_05 So, if Earth and other planets were involved in a hypothetical interplanetary basketball tournament, who is making Earth's starting 5?

[Lost Internet @LostMemeArchive]

NBA Jam had hidden code that made the Chicago Bulls miss last-second shots against the Detroit Pistons.

The creator was a Pistons fan.

So if the Bulls tried to win at the buzzer against Detroit, the game quietly sabotaged them.

Petty coding at an elite level.

[Mark Papadakis @markpapadakis]

@The_Pet_Way Real has no chance against that team without Tavares. Experience will count for nothing. Valencia with razzle-dazzle Real and they won’t know which way is north by the time they are done with them. For better or worse.

[Mark Papadakis @markpapadakis]

@philthistweet Very cool. BTW, case WM_CREATE: is missing a break and should fall through to case WM_COMMAND

[Mark Papadakis @markpapadakis]

Offloading/outsourcing thinking and problems solving to AI will make us dumber down the road. Too many upsides to AI technologies, but not without tradeoffs we may or may not want to have made.

[Mark Papadakis @markpapadakis]

@The_Pet_Way What is this supposed to be ? Some kind go fashion accessory ? It’s not a sports bra got man boobs right ?

[Mark Papadakis @markpapadakis]

@lemire I appreciate those kinds of insights/specifics involving tools, processes, and choices of technologies and platforms. Thanks!

[Mark Papadakis @markpapadakis]

@mitchellh @jarredsumner I understand people don’t care for copilot. I love it. I use it to understand and explore codebases. I used it to understand ghostty codebase too. It’s fantastic for such use cases. My only gripe with GH is that it’s just slow (pages and responses).

[Curiosity @CuriosityonX]

This is incredible.

This machine is capable of cleaning up 100 million kg of plastic ocean waste, and as of 2025, it has already collected about 500,000 kg of plastic.

It aims to remove 90% of ocean plastic by 2040.

[European Commission @EU_Commission]

We are introducing EU Inc. To make building and growing a business across the EU faster, simpler, and smarter.

🔸 Start a company in less than 48 hours
🔸 No minimum capital requirement
🔸 Fully online and borderless

[Mitchell Hashimoto @mitchellh]

Ghostty 1.3 is now out! Scrollback search, native scrollbars, click-to-move cursor, rich clipboard copy, AppleScript, split drag/drop, Unicode 17 and international text improvements, massive performance improvements, and hundreds more changes. https://t.co/IMk3i6528t

[Mark Papadakis @markpapadakis]

@MParakhin C++20 coroutines were/are brilliant in terms of design and semantics. I would say the same about concepts; agreed about modules. What's your issue with coroutines ?

[Daniel Vogel @EpicVogel]

I spent the last few days pushing Anthropic's VLIW performance take-home kernel optimization challenge to its limits using Claude Code as an orchestrator.

The task: schedule operations for a custom VLIW SIMD architecture running a tree traversal with hashing. 256 items, 16 rounds, 5 execution engines with different slot limits.

Starting point: 147,734 cycles (naive)
Where Claude Code landed: 1,105 cycles — a 134x speedup

The journey involved hundreds of AI agents across several dozen iterations, exploring every angle: hash algebraic merges, L4 tree caching, DAG-based list schedulers, 250K+ parameter configurations, emission order sweeps, store engine exploitation, and loop-based kernels. The last gain was 2 cycles, found by sweeping 103,000 configurations.

I then had Claude write a formal lower bound proof in Lean showing the kernel cannot run in fewer than 1,081 cycles — proven from load engine capacity (2,089 ops at 2/cycle = 1,045 minimum) plus unavoidable dependency overhead.

For verification I relied on Anthropic's test harness and extended it with randomized parameter testing and also extended it with Kernel Optimization Fun's output index verification.

What struck me is how good Claude Code has gotten at orchestrating optimization work. It ran teams of 10 parallel agents in isolated worktrees, each exploring different hypotheses. Agents communicated findings, dead ends propagated instantly, and the system converged on proven optima. The DAG scheduler that broke through a 6-iteration plateau came from Codex (gpt-5.3) running through Claude Code's MCP integration — multi-model orchestration improving the solution.

Proebsting's Law says compiler optimizations double program speed every 18 years. AI agents with the right tools are compressing that timeline dramatically — not by improving compilers, but by doing the work compilers can't: reasoning about problem structure, exploring architectural trade-offs, and proving bounds. The 134x speedup here came from algorithmic insight (merged hash stages, path-bits scheduling, bias-free C5), not instruction selection.

Wild times for performance engineering.

[Mitchell Hashimoto @mitchellh]

Daydreaming about an entry MacBook Neo as a thin client for local LLMs and dev work on a maxed future Mac Studio. Apple is creating the best value hardware for local AI and it’s amazing to think of consequences of that.

[Mark Papadakis @markpapadakis]

One of my hobbies has always been studying codebases; AI/LLMs transformed that in so many ways -- it's so easy to just have them analyze codebases now, ask questions, and effectively compress the time it takes to grok things.

[Mark Papadakis @markpapadakis]

@drandakis That's pretty what the EU has been reduced to: coming up with restrictions and regulating anything and everything. The bureaucrats took over, and they do the only thing they knowhow to do: regulate. See also, https://t.co/Gb7d80nUBY

[Mark Papadakis @markpapadakis]

I just published The Cat’s Out of the Bag: Living With AI and Looking Ahead https://t.co/OKYkWK9hsR

[Mark Papadakis @markpapadakis]

@exQUIZitely Bard’s Tale 1 Skara Brae map is up there too

[Mark Papadakis @markpapadakis]

@LunarMinxDev This looks fantastic. I myself care for exploration, traps, puzzles - not very punishing or elaborate(see FTL's Dungeon Master), or rare (see Bards' Tale 1). I love the aesthetics.

[Mark Papadakis @markpapadakis]

@aripap The whole notion of tips ( other than tipping waiters and such ) makes little sense to us Europeans, but you go ahead. Reinforce that AI with your monetary feedback.:)