Flo 🔶
@faz.ms
Interested in #Infosec and #ProductiveDisagreement | #StayAtHomeDad, worked in #Biotech, co-built tiny companies in renewables and structural engineering sectors, ex #HumanRights observer | likes #ElixirLang, #Boardgames |🔸 #10PercentPledge
Once MCP had reached market saturation and every company had connected their Slack and Google workspace to Claude, Anthropic upgraded their internal Clio tool to Claude‘s Capitalist Coordination Protocol, finally giving the world an actual working Cybersyn.
November 3, 2025 at 9:22 PM
Once MCP had reached market saturation and every company had connected their Slack and Google workspace to Claude, Anthropic upgraded their internal Clio tool to Claude‘s Capitalist Coordination Protocol, finally giving the world an actual working Cybersyn.
And yes, Turing also anticipated the connection to ESP mentioned in the last paragraph of the article, philosophizing about how psychokinetic powers might affect random number generators
November 2, 2025 at 9:45 PM
And yes, Turing also anticipated the connection to ESP mentioned in the last paragraph of the article, philosophizing about how psychokinetic powers might affect random number generators
Surely as someone who’s written 121,000 words about what’s wrong with society, he’s been posting the same image just with the spear pointing to the right for every mosque, synagogue, supermarket and other right wing inspired shooting in the last decade.
September 11, 2025 at 9:56 PM
Surely as someone who’s written 121,000 words about what’s wrong with society, he’s been posting the same image just with the spear pointing to the right for every mosque, synagogue, supermarket and other right wing inspired shooting in the last decade.
Always fun and unsettling at the same time seeing LLMs come to terms with the fact that we live in an out-of-distribution world compared to their training data.
August 28, 2025 at 10:47 PM
Always fun and unsettling at the same time seeing LLMs come to terms with the fact that we live in an out-of-distribution world compared to their training data.
August 2, 2025 at 9:39 AM
claude code custom subagents are amazing, both to get better results and to avoid hitting your daily limits by outsourcing to Gemini.
July 30, 2025 at 10:38 AM
claude code custom subagents are amazing, both to get better results and to avoid hitting your daily limits by outsourcing to Gemini.
The amount of hours I've spent frustratedly chasing bugs which turned out to be "mixed up _ and -" on it's own already justifies paying for claude code.
And yes I do use linters in every language, and no they don't catch the trickier versions of these.
And yes I do use linters in every language, and no they don't catch the trickier versions of these.
July 14, 2025 at 9:10 PM
The amount of hours I've spent frustratedly chasing bugs which turned out to be "mixed up _ and -" on it's own already justifies paying for claude code.
And yes I do use linters in every language, and no they don't catch the trickier versions of these.
And yes I do use linters in every language, and no they don't catch the trickier versions of these.
For merge conflicts I now actually like to let Claude code use gemini-cli to outsource the token-intensive busywork to Gemini 2.5 pro and it's large context window. That way I don't reach my Opus quota as quickly.
![Terminal screenshot showing git merge conflict analysis for Phoenix LiveView files. Text reads: "Bash(cat lib/crux_hub_web/live/statement_live/view.ex [...] | gemini -p 'I have merge conflicts in 5 Phoenix LiveView files. Please analyze each file's conflicts and tell me: 1) Which conflicts are just formatting/syntax differences (HEX vs HEEx, indentation, etc.) that I should resolve by accepting incoming version, and 2) which conflicts involve actual logic changes that need manual resolution. The files are from a feat/user-profile branch being merged into main. Be very concise in your analysis.') Based on my analysis of the merge conflicts, here is a summary for each file: ### 1. 'CruxHubWeb.StatementLive.View' * ++Formatting/Syntax Conflicts (Accept incoming 'feat/user-profile' version):++ * Multiple instances of added newlines and indentation for readability. * Migration from EEx ('<%= ... %>') to HEEx ('{...}') syntax in the template. [...] ### 4. 'CruxHubWeb.UserAuthTest' * ++Formatting/Syntax Conflicts (Accept incoming 'feat/user-profile' version):++ * All conflicts in this file are purely stylistic, involving added newlines and indentation to improve code formatting. * ++Logic Conflicts:++ * None. using macos seatbelt (profile: permissive-open) ... 22:51 Perfect! Based on Gemini's analysis, I can see that most conflicts are formatting/syntax differences that should accept the incoming version, with only a couple of logic conflicts that need manual attention. Let me handle them: Bash(git checkout --theirs test/crux_hub_web/live/user_live_test.exs [...] && git add test/crux_hub_web/live/user_live_test.exs [...]) Running.."](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:huocg7zsbthk54vjai56wd27/bafkreihtklkoosrexxyutamk3tvd2xkx7jf5cct4lt2y5qx3dd3bj4zvdm@jpeg)
July 6, 2025 at 9:01 PM
For merge conflicts I now actually like to let Claude code use gemini-cli to outsource the token-intensive busywork to Gemini 2.5 pro and it's large context window. That way I don't reach my Opus quota as quickly.
Nope, I didn’t mean it that way, you didn’t hurt me, and I felt our exchange was completely friendly.
It’s just that I sometimes post about AI, and that has landed me on quite a few blocklists and it’s sometimes difficult for me to judge when my questions come off as annoying
It’s just that I sometimes post about AI, and that has landed me on quite a few blocklists and it’s sometimes difficult for me to judge when my questions come off as annoying
June 26, 2025 at 5:57 PM
Nope, I didn’t mean it that way, you didn’t hurt me, and I felt our exchange was completely friendly.
It’s just that I sometimes post about AI, and that has landed me on quite a few blocklists and it’s sometimes difficult for me to judge when my questions come off as annoying
It’s just that I sometimes post about AI, and that has landed me on quite a few blocklists and it’s sometimes difficult for me to judge when my questions come off as annoying
they ask you how you are, and you just have to say that you're fine, when you're not really fine, but you just can't get into it because they would just press esc
June 24, 2025 at 9:11 AM
they ask you how you are, and you just have to say that you're fine, when you're not really fine, but you just can't get into it because they would just press esc
Dear #elixir people: what's the best way to mock streams in tests? I landed on this solution, but I wonder if there is something simpler I'm missing?
June 24, 2025 at 9:02 AM
Dear #elixir people: what's the best way to mock streams in tests? I landed on this solution, but I wonder if there is something simpler I'm missing?
