engineer | conjuring @webuildthingsx | tips on using & building with agents

Joined March 2024
220 Photos and videos
Pinned Tweet
Claude Fable 5 one shot build. Can you collect all the orbs in given time limit? (fog is intentional)
8
1
15
1,214
harsh retweeted
Jun 13
don’t let anyone stop you from going on THAT trip of your life!
21
1
186
5,596
Built a solar system simulation with Claude Fable 5. Includes: > JPL ephemeris for all 8 planets. > 56 asteroids from JPL/MPC. > 12K belt particles with actual Kirkwood gaps. > Apophis 2029 flies past Earth on the right date. One shot Result 👇
1
3
5
273
Didn't saw this one coming! Apparently @claudeai fable 5 is so cracked , that it can fix my breathing issues using code. Amazing progress towards AGI. Jokes aside: Dear claude team & @bcherny, if you see this , kindly tune your responses to generic human cases as well.
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use. Its capabilities exceed those of any model we’ve ever made generally available.
2
2
4
434
Someone please tell me how to activate sandbox environment in codex in windows. Every time I do it , it says ( in simple words) "Couldn't run code in sandbox , because windows runtime failed" Not even able to use the inbuilt browser. Are Window users doomed ? @OpenAIDevs
98
harsh retweeted
People think numbers are "things you can calculate." Actually almost every real number is so complex that: no algorithm can generate its digits, no finite description can specify it, no human will ever reference it. The numbers you know (π, e, √2) are an unimaginably tiny special subset.
19
7
82
6,832
harsh retweeted
Been obsessed with how much data exists inside an F1 race ! 😛 built "Chronos F1" : #F1 #AI #IBM
2
6
217
Get on the ship to ship!
We’ve decided to gather together the best builders in the bay… on the bay. Our 89-foot yacht will bring together the 30 best applicants to have them battle it out as we sail out under the bridge for a final sunset demo. Is this a good idea? A terrible one? Who knows. But @composio along with @nebiusai and @tavilyai is determined to find out. And the winner is taking home a DGX Spark and $50k in credits. Apply by posting what you’ve built and tagging @ship_builders, @composio, @nebiusai, @tavilyai
3
92
Psychology says the people who can spend an entire afternoon doing something with no measurable output aren’t being unproductive, they’re maintaining the only quiet wing of consciousness where new ideas have ever actually been built! Credits : Leravi
3
86
harsh retweeted
Most people try to reduce Claude Code costs by reducing tokens. That's the wrong optimization. The real bottleneck is context quality. When context gets noisy, performance drops long before the context window can even fill up. Here are 8 ways to I fix this. A thread 🧵
3
3
14
906
Most people try to reduce Claude Code costs by reducing tokens. That's the wrong optimization. The real bottleneck is context quality. When context gets noisy, performance drops long before the context window can even fill up. Here are 8 ways to I fix this. A thread 🧵
3
3
14
906
The biggest shift is this: Stop treating context like a storage. Start treating it like working memory. The best Claude Code users aren't feeding it more context. They're feeding it better context. Check out the complementary blog in comment, for details, if curious Thanks
1
1
117
1
77