"LOOK AT THIS SHOT!"
I can't sing the praises of our home SNY crew loud enough. Cameras nail the definitive angle while also capturing the emotion and suspense of the moment.
+1 for "context engineering" over "prompt engineering".
People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window with just the right information for the next step. Science because doing this right involves task descriptions and explanations, few shot examples, RAG, related (possibly multimodal) data, tools, state and history, compacting... Too little or of the wrong form and the LLM doesn't have the right context for optimal performance. Too much or too irrelevant and the LLM costs might go up and performance might come down. Doing this well is highly non-trivial. And art because of the guiding intuition around LLM psychology of people spirits.
On top of context engineering itself, an LLM app has to:
- break up problems just right into control flows
- pack the context windows just right
- dispatch calls to LLMs of the right kind and capability
- handle generation-verification UIUX flows
- a lot more - guardrails, security, evals, parallelism, prefetching, ...
So context engineering is just one small piece of an emerging thick layer of non-trivial software that coordinates individual LLM calls (and a lot more) into full LLM apps. The term "ChatGPT wrapper" is tired and really, really wrong.
The more C++ I have done, the less confident I feel in my understanding of the language. I finally wrote down my experience with compiler bugs that are kinda symbolic of my issues https://t.co/l9W0sd8gaa
@CDisillusion This was a blog post that covered how they decoded text that was pixelated instead of blurred by @bishopfox https://t.co/GLczIjQn92
While they didn't tackle blurred text, you can kinda see how their strategy could be extended. Just doesn't seem worth the risk
@jakubtomsu_ I think it's fair to say that the average developer should trust the compilers. Hand optimizing is going to be a waste for the average developer because they are going to be applying microopts that don't matter. If you are familiar with profiling tools then you can do it right
@jetset_ It's a great point! It's also why a lot of archivists talk about video games as a cultural hole because so many important games or phases of games just disappear as servers shut down
I recently started skateboarding and was curious to find how many other people shared my skateboarding stance. So pulled some data and did some analysis! https://t.co/LidmYojE4V
TL;DR: There are many right-foot dominant people riding goofy!
#skateboarding#goofy
Our mission to explore Jupiter's moon Europa remains on track, with a launch period opening on Oct. 10. We recently conducted extensive testing of our spacecraft's transistors, and results suggest they're ready to support our baseline mission. https://t.co/M4hClJroBB
@bwesterb Is there a reason security semantics don't get added to the standards? Isn't that the right way of stopping compilers from breaking secure software?