I primarily used Claude Code to build web apps. Today, I tried Codex for the first time and asked it to redesign the landing page. The results were incredible!
Many people believe *most developers* use Claude Code or similar to write their code. This place is a bubble. Go outside and most developers have never even heard about Claude Code and are using ChatGPT to write dinner recipes.
Introducing Claude Haiku 4.5. Our latest small model that matches Sonnet 4's performance at a third of the cost and more than twice the speed.
Introducing Claude Haiku 4.5: our latest small model. Five months ago, Claude Sonnet 4 was state-of-the-art. Today, Haiku 4.5 matches its coding performance at one-third the cost and more than twice the speed.
Early results for Claude Haiku 4.5 are strong: 73.3% SWE-bench Verified (coding), claiming Sonnet 4-level coding at one-third the cost and over twice as fast; priced at $1 in / $5 out per million tokens. Safety system card: ASL-2, with lower misalignment than Sonnet 4.5 and Opus…
We reduced wrong-domain search results from 12% to under 5% by treating queries differently. Most RAG systems use uniform scoring, which is flawed. Here's what works: **Intent Matters** - "What is accounts payable?" → Definition query Needs: Conceptual depth, established…
How long it normally takes to have any followers in X? I have been grinding for a week now...
BREAKING: NVIDIA CEO Jensen Huang hand-delivered the world’s smallest AI supercomputer to Elon Musk.
Grok voice mode is insanely advanced It pulls real-time news and information straight from the 𝕏 live feed, combines this with instant web search and source citations, offering capabilities no other AI can match
Sam Altman says his personal benchmark for advanced AI is when it can do science. With GPT-5, he sees the first sparks, AI making math discoveries or aiding biology research. These early examples are small but meaningful signs of what's coming. He believes that in two years,…
One bad idea can waste a year of your life as a founder. The real skill isn’t spotting “genius” ideas. It’s avoiding the bad ones that look good on the surface. Avoiding a bad SaaS idea >>>> Finding a good SaaS idea.
Google Gemini 3 created a clone of the Windows operating system, all in a single shot. It seems to really excel with AI coding. The public release of Gemini 3.0 Pro is expected for this month.
Just type “add a girlfriend” to any video on the new Grok Imagine
Steve saw the future as a bright and boundless place, lit the path forward, and inspired us to follow. We miss you, my friend.
Quick experiment: When I ask LLMs to "think from first principles," responses are sharper, with less pattern matching and more reasoning. Anyone else tried this? What prompts worked for you?
United States 趋势
- 1. #TORQSports N/A
- 2. Russ 19K posts
- 3. Argentina 510K posts
- 4. Waddle 3,791 posts
- 5. Rickey 2,401 posts
- 6. Big Balls 24.9K posts
- 7. $HIMS 4,302 posts
- 8. Olave 3,228 posts
- 9. #BeyondTheGates 5,774 posts
- 10. Malcolm Brogdon N/A
- 11. Aphrodite 4,945 posts
- 12. #ClockTower1Year N/A
- 13. Kings 164K posts
- 14. Voting Rights Act 29.3K posts
- 15. #TrumpsShutdownDragsOn 7,123 posts
- 16. Capitol Police 28.1K posts
- 17. Maybe in California N/A
- 18. Meyers 2,458 posts
- 19. annabeth 2,678 posts
- 20. Justice Jackson 21.1K posts
Something went wrong.
Something went wrong.