Many countries across the world provide emergency preparedness manuals to all citizens. But I understand it’s a big ask to expect it in India.
They can however start with training police and highway staff - who are first responders - to guide people in these situations using SOPs. A fire in a tunnel is not a very unique scenario, is it?
Even if that’s too much to expect, install fire suppressing systems in these tunnels and divert traffic away from it.
Ten years of walking around cities catching Pikachu also built one of the largest urban image datasets on Earth.
It's now training delivery robots to navigate sidewalks.
Thanks a lot Pokemon Go.
I had some vibes that Opus 4.8 was performing worse than older ones for some of uses that are off distribution and now I have the receipts. Latest Opus/Sonnet are causing tool invocation failures on Pi's edit tool when older ones did not! I wrote about it. https://t.co/96Qd8zfdVc
One thing I do not like about prompts is how we hardcode them in code. Every prompt change leads to a code change, followed by a redeployment.
Hence, today I am launching px0 - an open-source prompt infrastructure toolkit that lets you version, update, and govern prompts in production, eliminating the need to hardcode prompts or redeploy your application.
It is ready to use with a complete API suite, management console, and SDKs in languages that we typically use to prompt LLMs - Python, Node.js, and Go. Here is a gist of the benefits
- no hardcoded prompt strings
- deploy-free prompt updates
- 5-line SDK integration - py, js, and go
- prompt versioning and lifecycle management
- self-host in your own cloud or VPC
- fully open source under the MIT license
Also, many more features are coming out in the next few weeks.
If you decide to give it a shot (link below) and run into any issues or have feature ideas, feel free to open an issue or a feature request. I would love to hear.
If you like the project and find it useful, please star the repo and show some love :) It would mean a ton.
Amongst my friends, Spotify is the lowest quality consumer app we still pay for. It certainly hasnt gotten noticeably better in the last couple years (arguably worse). So, this is not the positive look Ant and Spotify are spinning here.
Bigger picture, this is the problem with a lot of AI reporting. It reports completely meaningless metrics like deploys per day or LoC. Why don’t we start reporting consumer satisfaction reports? Actually end state research results.
All the no nuance AI people always come out and think that this is anti AI. Again, I think AI is great and Claude is great. But this is bad marketing and makes both look like clowns.