@simonw

–

10 April 2024 3h52

Who will be the first LLM provider to publish genuinely useful release notes?

Overcoming AI Output Token Limits Through Strategic Prompting

By

–

10 April 2024 3h33

Yeah the maximum number of output tokens is an important factor – but you can overcome that by promoting it for more content and including the content it's already given you

Why Transparent Remote Calls Failed in Distributed Computing

By

–

10 April 2024 2h41

I thought we gave up on SOAP and CORBA and DCOM and the like precisely because it turns out completely transparent method calls and object serialization are more trouble than they are worth

Serialization formats for network object communication

By

–

10 April 2024 2h39

Maybe there are versions of this that work really well in other languages that I don't know about, but the impression I have is that serializing objects for network communication is much better handled using formats like protobufs or msgpack or even JSON

Runtime Class Implementation Identity Verification Across Connections

By

–

10 April 2024 2h38

How would you confirm at runtime that the class implementations on either side of the connection were indeed identical, including any code that they depend on?

Gemini Pro 1M Token Context Now Free Without Waitlist

By

–

10 April 2024 2h28

Gemini Pro with the 1 million token context limit is now available for free and without a waitlist! https://t.co/eiNsXCBMBU
— Simon Willison (@simonw) 10 avril 2024

Gemini Pro with the 1 million token context limit is now available for free and without a waitlist!

Discovering Prompts Behind AI Model Research Paper

By

–

10 April 2024 2h16

I'm trying to figure out what prompts they used. I think it's these – this is from a paper that their paper links to https://
github.com/lilakk/BooookS
core/tree/main/prompts
…

Fiction vs Non-Fiction in AI Summary Evaluation Studies

By

–

10 April 2024 2h13

Looks like all but one of the books in that study were fiction, which I imagine has quite different characteristics when it comes to evaluating summaries than non-fiction https://
github.com/mungg/FABLES/b
lob/main/booklist.md
…

Google Abandons Infrastructure Service Amid Cloud Team Concerns

By

–

10 April 2024 2h04

Did anyone at Google think to ask the Google Cloud team how they felt about having yet another high profile instance of Google abandoning an infrastructure service that tons of people used and loved?