Toast

Thread: The S&P 500 on average returns 7% a year, but when should you invest?

@vasuman this kind of click bait posts calls for a block. don’t wish to see this trash on my feed. suggest everyone blocks it as well

Claude Mythos just refactored my entire codebase in one call. 25 tool invocations. 3,000+ new lines. 12 brand new files. It modularized everything. Broke up monoliths. Cleaned up spaghetti. It worked.

what happened to lsof -i :3000 -P -n ? not sure why we need a cli for this

@itsolelehmann would focus more on multi-models instead of multi-role and also i would make it more a decision loop focus. Hypothesis -> Council -> Experiment -> Feedback(Council runs again) that way you have data that you actually work with and it becomes a decision thinking engine

@wesbos use a GPG key

scam. They are spoofing the commits on github with my email. How do I stop this? I guess I should be signing my commits?

Their DevSecOps became DevSecOops. HugOps to the person who pushed this.

@geezaz_ @witcheer @bootdotdev it’s worth it. it’s like starting a jigsaw puzzle, you can’t see how it all fits at first but suddenly all the pieces will start making sense and you’re better off than the vibe coder because you understand systems.

Inspired by @witcheer's progress, I’ve officially started my own coding journey with @bootdotdev! I already feel pretty dumb. is that normal? I’m a total beginner, but my biggest worry is that I’m historically terrible at math. Is it still worth pushing through, or am I going to hit a wall? Would love any advice or encouragement for a math-challenged newbie! 🤟

@thetripathi58 just brew install mole and fix it. don’t need a thread on this.

If Apple says your Mac disk is almost full and you need to buy a new laptop, do this first. I went from 249/256GB to 110GB without deleting a single active project. Here is the exact playbook:

@ashwinhegde19 @bcherny the issue with this is the cache.. it’s more costly to switch models halfway.. you should plan in one and execute in another. hope this helps!

@bcherny still no easy model switching mid-session without killing context?

I wanted to share a bunch of my favorite hidden and under-utilized features in Claude Code. I'll focus on the ones I use the most. Here goes.

@bcherny can’t wait for all of these hidden features to work with bedrock as well ☺️

@daemonxbt if you come for the king, you best don’t miss.

Zachxbt in 2024: I found who stole $240M Zachxbt in 2026: I found 10 fake twitter accounts

@penguinpecker1 @star_okx i think you should read the post before commenting. 😂😂😂

@star_okx what took you so long to solve this, did he needed to file complaint with authorities?

Dear Anizja, We’re sorry to hear about your experience and want to assure you that we are actively working on your case. We would also like to clarify a few key points: 1.BTC does not operate on the TON network and cannot be deposited via TON. 2.The asset sent to the TON network is a different token labeled “BTC”, not native Bitcoin. 3.This TON-based “BTC” has been flagged as a scam token by multiple third-party blockchain explorers. 4.Nevertheless, we are actively working to assist in recovering the asset. As this involves a scam-labeled token, it falls outside standard procedures and may require additional time. We kindly remind users to remain vigilant and exercise caution when interacting with scam tokens. Thank you for your patience and understanding

@VirtualKenji you might want to look at adding qmd as a posthook or a cronjob that runs daily for the rerank.

Your Claude keeps forgetting shit? Not a memory problem, but a retrieval problem. I've been obsessed with building a content system on Claude that writes by itself and learns from its mistakes. 250 sessions, 126 error files, 65 lesson files, 16 GitHub repos analyzed, and a daemon (auto-writer) that writes tweet drafts at 2 AM. Last night the daemon wrote tweets bragging about how I put auto-fire rules into CLAUDE md. The problem: This was the exact thing I spent all day removing! My auto-writer had zero awareness of what I fixed this week... it made me look like a dumbass for shilling the (wrong) implementation I spent a whole day fixing. Root cause: my system had ONE retrieval method. Glob for file names. Grep for text. That's it. My semantic search tool (qmd — hybrid BM25 + vector embeddings + LLM reranking) had been stale for a month. 675 of 5,058 files indexed. 13% visibility. The most powerful search tool in my vault wasn't firing. The fix is a router, not a waterfall: Query type → tool Known file path → Read directly File name pattern → Glob Exact text match → Grep "What's relevant to X?" → qmd vsearch Cross-domain patterns → qmd query (full pipeline with reranking) Open-ended exploration → Agent subprocess No waterfall. No "try this first, then fall back." Match the tool to the query type. Period.

if you’re on a mac, cmux is the way to go.

and i haven’t started on my actual work tickets. 😅 but it’s supposed to help me work faster so it’s okay i guess..?

Started the day building a Claude Code skill that could spawn its own instances via cmux… ended up creating a full-fledged go-lang CLI tool for memory and session management via MCP as well.

@VirtualKenji I’m building with Claude Code every day and sharing what I learn along the way. Not an expert! still figuring things out but the stuff I ship has to hold up in production, including compliance and security checks. So just sharing little tips and tricks i find useful ☺️

@Toastonomics bro I love how you randomly drop alpha that is super super useful Yes I will do this Thank you so much bro 🤝

"Just write clearer instructions." No. I wrote the clearest instruction possible: "Before any build task, search your error database first." Claude read it. Every session. For weeks. Never followed it once. > write clearer rule > Claude reads it > Claude ignores it > rewrite the rule > Claude reads it again > Claude ignores it again Sound familiar? The instruction was perfect. The ENFORCEMENT was missing. Same rule, but from "advisory" to "mandatory gate." Gates have triggers. Gates block actions. Advisory rules are "suggestions" you wish were instructions. Now it fires every time. Fuck "write it clearer." The enforcement mechanism trigger IS the rule.

@levelsio check out cmux buddy.

Apparently the tmux that does NOT suck exists It has an absolutely unpronounceable name and is called Zellij I will try it now thanks everyone for recommending it 😊 zellij.dev/about/

@witcheer a good refresher article for me 🙂‍↕️ good work. this should go viral