Arthur Reynolds
606 posts
Arthur Reynolds
@ArthurReyn
Occasional shitposter. Like AI, astronomy, software, math, the usual nerdy stuff essentially.
Barcelona, Spain Beigetreten Nisan 2020
36 Folgt39 Follower
People are dumb and taking this as a negative post when i'm just pointing out they're going for a darker tone earlier on which is refreshing
English
@MaMoMVPY Why would a prediction machine have zero intelligence?
English
No Claude do not have an IQ of anything. It is a prediction machine. A prediction machine trained on more data now than 3 years ago. But that's all. It has ZERO intelligence.
Charbel-Raphael@CRSegerie
In March 2023, Claude had an estimated IQ of 64. Today, Claude Opus 4.6 scores 133 on the Mensa Norway test. GPT-5.2 Thinking hits 141. Gemini 3 Pro, 142. That's a jump from cognitively impaired to gifted in three years. No human population has ever improved that fast, the Flynn effect gives us ~3 IQ points per decade. AI just did 70 points in 36 months.
English
call claude with dangerously skip permissions in a shell loop for example
echo "=== Phase 2: Fixing bugs ==="
MAX_ITER2=200
ITER2=0
while true; do
ITER2=$((ITER2 + 1))
if [ "$ITER2" -gt "$MAX_ITER2" ]; then
echo "Phase 2: max iterations ($MAX_ITER2) reached, breaking."
break
fi
ISSUE_NUM=$(gh issue list --state open --label bug --json number \
--jq 'sort_by(.number) | .[0].number')
if [ -z "$ISSUE_NUM" ] || [ "$ISSUE_NUM" = "null" ]; then
echo "No more open bug issues."
break
fi
echo "--- Fixing issue #$ISSUE_NUM ($ITER2/$MAX_ITER2) ---"
claude -p --dangerously-skip-permissions --model opus --effort max \
"Fix GitHub issue #$ISSUE_NUM.
First read the issue: gh issue view $ISSUE_NUM.
Fix the bug described. Run 'go build ./...' and 'go test -short ./...' to verify.
If both pass, commit with message 'fixes #$ISSUE_NUM' and push to origin.
Do NOT add Co-Authored-By lines to the commit message.
If the fix breaks build or tests, revert your changes and close the issue:
gh issue close $ISSUE_NUM -c 'False positive — fix breaks build/tests'."
English
Claude code rarely runs for longer than 15m without stopping and asking for input from me. How do all these stories of people letting agents run overnight work? Custom harnesses? Yelling at Claude in all caps to keep going no matter what?
English
@thsottiaux I don't like it. I explicitly mention it in CLAUDE.md to not add coauthored by in commits.
English
Do people like this? We don't do this for codex because it exists to help you and it's important that you remain the owner and accountable for your work without AI taking credit. At the same time it does mean that you can't trace how popular codex is among repos.
Yuchen Jin@Yuchenj_UW
I noticed something interesting: Claude Code auto-adds itself as a co-author on every git commit. Codex doesn’t. That’s why you see Claude everywhere on GitHub, but not Codex. I wonder why OpenAI is not doing that. Feels like an obvious branding strategy OpenAI is skipping.
English
Nobody wants to read AI-generated posts.
Nobody wants to read AI-generated emails.
Nobody wants to read AI-generated newsletters.
Nobody wants to read AI-generated reports.
Nobody wants to read AI-generated articles.
Nobody wants to read AI-generated anything.
Luiza Jarovsky, PhD@LuizaJarovsky
Nobody wants to read AI-generated books.
English
If we’re living in a simulation, who created it? What are they using us for?
English
@drtod1000 @haha_girrrl I've never heard your surname before
English
@haha_girrrl Theoretical physics is not a good area because you can ONLY theorize and work in academia pontificating, but you can't build a damn thing. Do Applied Physics and engineering and add machining to your skill set. You become valuable because you can now build what you imagine.
English
English
@garrytan Opus 4.6 with 1M tokens is dumb currently. I've been experiencing significant model degradation. I've now switched back to the 200k model.
English
I underestimated how powerful Opus 4.6 with 1M tokens is. Even last year we were absolutely hitting context limit problems constantly. 1M tokens means you can do much more complex analysis entirely in context. Claude Code is so much better.
This is the worst it will ever be.
English
@elonmusk It's less about "looking"/"observing", and more about the fact that in order to observe, we need to measure and in the process of measurement, we end up altering the state of the thing we are observing. Many people misunderstand that it's like a read-only operation on a database.
English
I had dinner once with a top physicist and a top computer scientist and asked what they thought the probability was that we were in a simulation.
They answered simultaneously at 0% and 100% respectively. It was like a double-slit experiment, but with humans.
Interstellar@InterstellarUAP
🚨 Simulation Theory: The Double Slit Experiment proves particles act like waves until observed then they snap into particles. What if our reality only "renders" when we're looking, just like a video game optimizing resources? Check out this episode from The Why Files breaking it down, tying it to Simulation Theory. Are we in a sim? This could be the key to unlocking the true nature of existence! The Why Files video did a great job on explaining the Double Slit Experiment & Simulation Theory What do YOU think—real or rendered? Drop your thoughts below!
English
why don't we have famous physicists anymore?
Andrew Côté@Andercot
why don't we have famous physicists anymore?
English
@RealestMemes_ Exactly, and the word derives from Latin pusillus ("very small") and animus ("spirit"), essentially meaning "very small spirit".
English
I hope not
I can’t stand Slack
Ivan Burazin@ivanburazin
Agents won't vibe-code a new Slack every time they need to communicate. They'll use the same Slack because the other agent's team also uses Slack. Traditional SaaS will survive because standards don't get disrupted by probabilistic code generation. Network effects and standardization still matter.
English
@CodeByNZ orders endpoint should be in a different file, and why does it have both team-members and employees; please take this image away from my eyes, I can't look at this
English
Baghdad looking good in 2026?
Clash Report@clashreport
Outdoor dining, Shahed and C-RAM. 📍 Baghdad, Iraq.
English
English
@MRowean @felixrieseberg You can already do this with code via remote-control
English
We're shipping a new feature in Claude Cowork as a research preview that I'm excited about: Dispatch!
One persistent conversation with Claude that runs on your computer. Message it from your phone. Come back to finished work.
To try it out, download Claude Desktop, then pair your phone.
English
I think it works best when you start from pre-existing code that follows good design patterns (controller, service, repository).
Some popular kits:
Quetic Basic Setup (Gumroad) - Go + Vue.js + PostgreSQL (I personally use this because this is the tech stack I like)
adityarise24.gumroad.com/l/basic-setup-…
Open SaaS (GitHub, free) - React, Node.js, Prisma
github.com/wasp-lang/open…
English
does anyone have any tips on how to prompt/plan when trying to oneshot large projects, like 50K+ LOC?
English
we are 6 months away from AI replacing all software engineers
Kalshi@Kalshi
JUST IN: Job postings for software engineers on Indeed reach new 6-month high
English