내가 좋아할 만한 콘텐츠
After last week's eval-no-eval world war, let's go back to eval. My blog on Multilingual Cypher Evaluation is now live on Arize! But it’s more than just a read, follow along with the notebook, spin up an Arize account, or run Phoenix via Docker locally, and try it out yourself!…
💡 Your LLM might ace English queries… but what happens when your users switch to Spanish, Hindi, or Mandarin? For teams building global AI systems, this is the hidden challenge: LLMs often fail to generate correct Cypher queries across languages. That means gaps in reasoning,…
You can master your music lessons on @ArizePhoenix
Readers responded with both surprise and agreement last week when I wrote that the single biggest predictor of how rapidly a team makes progress building an AI agent lay in their ability to drive a disciplined process for evals (measuring the system’s performance) and error…
Those days when you had a small win and a big loss. Not in great mood. Had high hopes! But we march on!
AI killed the illusion of writing code is what made you a developer. Writing code is a very tini-tiny part of development. – can you break a messy problem into steps? – can you see edge cases? – can you decide which trade-off is worth making? AI didn’t kill developers. It…
75% of the software engineering is KTLO, Migrations, Incident handling, Documentation. 20% is coding. 5% is complaining that 75% is non-coding.
Requirements part. In real world, requirements are not leetcode solvables.
Optimize Cline on SWE Bench using Prompt Learning and see improvement for yourself! Code: github.com/Arize-ai/promp… Use the Prompt Learning SDK: github.com/Arize-ai/promp… Use Phoenix to run: LLM Evals: arize.com/docs/phoenix/e… Experiments: arize.com/docs/phoenix/d…
Why databricks is there?!
Ok so cancer is out of distribution.
Good question, it's basically entirely hand-written (with tab autocomplete). I tried to use claude/codex agents a few times but they just didn't work well enough at all and net unhelpful, possibly the repo is too far off the data distribution.
Live Translated by @SarvamAI
#Watch | LG Electronics India MD Hong Ju Jeon Delivers Entire Speech in Hindi at NSE Listing Ceremony #LGElectronics #India #NSEListing
- No human
Things that will become more common in Indian households: - Cleaning Robots - Air Conditioners - High Quality Cameras - Paintings - Smart Cookers - Electric Vehicles - Voice Controlled Lighting - Smart Locks Interesting 5 years ahead for builders of these
> I am not shipping code I don’t understand. World would be a better place if more engineers took this to heart ❤️
Ran some deep research on data-oriented design, which dictates your data structures according to hardware reality and not some worldly abstractions. blog.aximox.com/data-oriented-…
This is getting worse and worse with the new engineers. No one is interested in reading documentation, working on refining requirements, generating technical specs where needed. Craft is lost.
AI is accelerating product bloat. Teams can ship features 10x faster, so they do - without asking if they should. Result: Frankenstein products with features bolted on, no coherent design, high cognitive load. AI gives you 80%. Your job is the remaining 20%: solve real user…
Just a matter of time.
About 2-3M people in various small to big IT service companies makes up for the consuming middle class across the big 4 cities in India (DEL MUM BLR HYD) The consumer discretionary sector & finance is keeping the India story ticking. Lot of pain if this domino falls over 🥲
United States 트렌드
- 1. Good Friday 45.5K posts
- 2. #FridayVibes 4,298 posts
- 3. #FanCashDropPromotion N/A
- 4. Dorado 4,247 posts
- 5. Mamdani 247K posts
- 6. Cuomo 102K posts
- 7. Flacco 97.6K posts
- 8. Happy Friyay N/A
- 9. Red Friday 1,977 posts
- 10. Justice 338K posts
- 11. #Talus_Labs N/A
- 12. Finally Friday 2,157 posts
- 13. Melly 3,730 posts
- 14. #FridayMotivation 3,066 posts
- 15. #FursuitFriday 11.3K posts
- 16. Pence 91.6K posts
- 17. Arc Raiders 3,232 posts
- 18. Bolton 246K posts
- 19. Tomlin 24.5K posts
- 20. Brandy 9,828 posts
Something went wrong.
Something went wrong.