← Feed

Posts

• Post #4456151

  Rule of thumb: if I count at least 1 instance of each these words/phrases in a post on Reddit then I bet is 99% chance its AI-gen and trying to trigger real humans into blooming crops (posting a reply comment) for their content farm ingest pipeline: * actually * genuinely * honestly * curious * legit * honest feedback * roast me * here's where it gets interesting * what I did not expect * (...many more...) one alone does not trip my spidey sense. if I see all of them it reeks of inauthenti...

• Post #4455954

  There are 3 types of people: 1. those who watched the new He-Man film in theatres 2. those who follow Skeletor on social media cc @skeletor@mas.to

• Post #4455763

  shell scripts FTW relying on vanilla Linux/UNIX features FTW rule of thumb: if it worked just fine 20-40 years ago it will still work today and in 20-40 years from now. not guaranteed, obvs, but a *wise* way to bet kids: get off my damn lawn, adults have real work to do

• Post #4430305

  tired: LLMs wired: libraries

• Post #4401501

  TODO self-host more TODO plain text blogging, and on site/host I control. (tho I have no plans to abandon Mastodon.)

• Post #4386037

  your hobbies: using LLMs, vibe codin and churning out AI slop, racing to a global bottom and devolving yourself into an interchangeable commodity painlessly replaced by bots owned by rich capitalists who are so wealthy to themselves already be post-scarcity and thus incented to act on every selfish sociopathic whim no matter how cruel or ultimately destructive to society, sustainable ecosystems or to a stable democracy itself mine: reading, thinking, brainstorming, writing, programming, (actual...

• Post #4348199

  I asked Google's AI to count to 60. just now THEY MESSED IT UP clowns. at scale. with billion dollar valuations. and non-stop lies, fraud and security risks, and massively wasteful resource usage #AI learned of this first from: https://fosstodon.org/@atoponce/117029422208891428

• Post #4255561

  I wonder if grift.ai is taken?

• Post #3670236

  goose not valued at Microsoft

• Post #3670235

  I evaluated a FOSS CLI tool whose capabilities would be nice to add to my toolbox did a quick eval of their public GitHub repo among other things I looked at their list of top contributors their names/locations? in no particular order (or in exhausting completeness): Chinese, Chinese, Chinese, African, Indian, Indian, Russian, Russian, Russian, Russian. a few American-ish or Western European adult-ish males. several profiles of undeterminable age/gender/ethnicity/nationality/residence. (I m...

• Post #3670234

  The 80s and 90s are back again, because we got too far away from what they did right and have come to regret it: * terminals * BBS * small tools that did only one thing and nailed it * web rings, grassroots * product reviews & curation only by individuals you trust, not some rando BigTech/gov algo or foreign botfarm-for-hire * decentralization * p2p * offline as normal * zero surveillance * true free speech: no cancel culture and random OpinionPolice mob bullying online * lies, fraud, p...

• Post #3670233

  ssh BBS (in 2026) FTW #ssh #BBS #terminals #shells #TUI

• Post #3661121

  its becoming apparent to me that continuing to try to read tech topic groups on #Reddit will cause me more frustration and irritation than is worth whatever gain I might ever get from it. and the latter is very rare there anymore. sad

• Post #3580192

  To summarize my take on the merits of the AI hype wave in software engineering, in broadstrokes: If you suck: LLMs rock!!! If you rock: actually, LLMs suck. #AI #LLM

• Post #3563229

  Slartboz: demo of my WIP realtime Rogue-like set in a dystopic post-apocalyptic future of NorMerika in the terrible-wonderful year of 2100 CE witness the lava accident in the Dork Fortress! witness the massive zombie herd! witness the gigantic Soviet AI cybertank bristling with lazer cannons! in general: witness! https://youtu.be/KPVIq39HoUM?is=7gW_rv1w4pS67caX #Roguelike #Roguelikes

• Post #3503296

  one of my dreams is/was to make a text-only game that makes me a lot of money. enough to live on, anyway maybe there's a greater chance now than say a decade or so. because LLMs and conversational chat bots have shown there is at least *some* paying market for it. and my skill set bridges both programming and creative writing dunno

• Post #2449884

  @raganwald I've kinda always viewed the strategic purpose of HN is to serve as a kind "shiny object" lure by YC (VC) to draw in young impressionable software engineers, to help seed or staff startups that they can invest in -- buy cheap lottery tickets in. Thats the ends, rest is just a means.

• Post #2274980

  its never been clear to me whether vim is a training ground for mastering NetHack... or vice versa either way it is glorious! heh #vim #NetHack

• Post #1868181

  A young Russian male (and who was ostensibly living in Russia for college, and all implications thereof) is pitching the world that we should trust his text editor with read/exfil access rights to our critical source code, secrets, PII, hosts, etc because -- and I shit you not -- his editor can show "rainbow brackets" rather than merely the old fashioned kind of syntax highlighting we've already had for 40 years or so. *groan* This is like a one-shot exemplar of the state of secu...

• Post #1816501

  I'm pretty aggressive about only ever speaking the truth. I think its the most sustainable and scalable of policy to have in life. Two more things sit adjacent to that: 1. I hate liars and lies. So much so that I've once gone as far as doing rigorous professional work at national scale to help fight back against them. 2. I hate being forced into making precise claims about slippery and fundamentally unknowable things, or into having an ostensibly serious conversation about a r...

• Post #1683332

  Today is the last day we can live our lives non-optimally. At midnight precisely we must all go on diets, begin training for decathlons and work furiously towards curing cancer. Until then? Stuff face with donuts while lounging around in pajamas.

• Post #1598514

  I have a hard time reading public news reports anymore about "US national security" and taking them seriously (notice the scare quotes?) because... we have none. not anymore we lost it 2025 January 20 and might not ever get it back

• Post #1598513

  I should prob figure out a way to opt-out of that Rust coreutils rewrite. Uncomfortable allowing it onto my boxes.

• Post #1598512

  if you offer a buyout option to employees whose age+experience >= 70 it means you are engaging in age discrimination #Microsoft

• Post #1547219

  look at who the US-Iran war benefits and who it hurts rule of thumb: that tells you who The Individual works for answer been obvious as far back as 2016

• Post #1547218

  Americans not allowed to buy new routers made by foreigners overseas. But we are allowed to use software made and run by foreigners overseas. And folks in foreign overseas call centers are allowed access to our critical life/financial PII. GOT IT!

• Post #1547217

  relative wisdom of uses for growing global supply of clean/renewable energy: worst: AI data centers better: data centers better still: scientific/medical HPC data centers best: EVs & battery recharging

• Post #1547215

  I too am willing to take $5B cash from Amazon, in exchange for promising to buy $100B in services from them in the future. WHEN WILL THIS AI BUBBLE POP? *readies big bowl of buttery popcorn* https://techcrunch.com/2026/04/20/anthropic-takes-5b-from-amazon-and-pledges-100b-in-cloud-spending-in-return/

• Post #1547214

  https://foreignpolicy.com/2026/04/21/trump-iran-us-negotiations-diplomacy/

• Post #1547213

  I dont think the AI VC insiders will let the AI bubble pop until OpenAI and Anthropic IPOs once the early holders cash-in the countdown to the pop begins Mythos and the like is just a delaying tactic. I bet all the key executives know the writing on wall and all they can do is FUD/YOLO, gaslight and self-deal a little more to stretch it out. The fundamentals just arent there.