ukuina

18 ms·

361.
by ukuina 2y ago
Expandability is no longer an option with Apple Silicon.
362.
by ukuina 2y ago
Why would you prefer one over the other?
363.
by ukuina 2y ago
If you turn off Battery Optimization, yes. Apps can even draw over and control other apps. Assuming the LAM is just an app controlling AOSP in a VM server-side, this really should have just been an app with the LAM sending touch events back
364.
by ukuina 2y ago
Seconds in the system tray clock is back now.
365.
by ukuina 2y ago
App Store search is horrendous. I find it better to use alternative front-ends like appsliced.co
366.
by ukuina 2y ago
Yes, input context window is reduced to 16k characters (not tokens). https://help.kagi.com/kagi/ai/assistant.html#tips-and-tricks...
367.
by ukuina 2y ago
Tested this on Apple Silicon at 128k context, very poor results with random character hallucinations.
368.
by ukuina 2y ago
This is the most applicable part of the article: Strategies to improve LLM accuracy: Retry: We repeatedly invoke a model with the temperature set to zero, up to five times, if it fails the test cases provided with the problem description. R
369.
by ukuina 2y ago
Using the openai Python library?
370.
by ukuina 2y ago
> Idea generation. I used to spend a lot of time chatting with colleagues about a vague idea I had. “How could we check whether X is true?” A tool like ChatGPT can help you get started. If you ask how to design an experiment to check a g
371.
by ukuina 2y ago
It was very surprising that Apple did not strong-arm the EU by threatening to pull out of the area. The short-term loss of revenue from leaving the EU is nothing compared to the long-term decline this is going to cause in Apple's servi
372.
by ukuina 3y ago
So many models fail at basic reasoning. > What weighs more: a pound of feathers or a great british pound? > A pound of feathers and a Great British Pound weigh the same, which is one pound. It works if you add "Think step by step
373.
by ukuina 3y ago
Why is LMsys flawed? Many people treat LMsys as gospel because it's the only large-scale, up-to-date qualitative benchmark. All the numeric benchmarks seem to miss real-world applicability.
374.
by ukuina 3y ago
The Xbox 360 with Gamerscore gained through Achievements launched in 2005. Lilely the first globally-synchronized and overt system designed for gamification.
375.
by ukuina 3y ago
What is Occam search? Can't seem to find it.
376.
by ukuina 3y ago
Please do not post replies that are trivial spelling or grammatical corrections; they add little to the discourse. To parent: I think all datastores can boil down to a SQL interface eventually. Someday a SQL-interface may dynamically choose
377.
by ukuina 3y ago
Kate is awesome and has been my default editor on KDE for decades. Pleasantly surprised it is available elsewhere!
378.
by ukuina 3y ago
If your tests pass, why not?
379.
by ukuina 3y ago
Have you tried response_format=json_object? I had better luck with function-calling to get a structured response, but it is more limiting than just getting a JSON body.
380.
by ukuina 3y ago
Preventing fast lanes will likely hamper the development of novel applications that we cannot even conceive of today. Banning throttling while keeping speedups open (subject to review) seems like a good balance.
381.
by ukuina 3y ago
Assuming that's going to be the datasheet naming.
382.
by ukuina 3y ago
Why the sarcasm?
383.
by ukuina 3y ago
This paper is specifically disproving the efficacy of agentic frameworks like AutoGen. Also, the built-in function-calling in GPT4 is simpler to use than AutoGen2's abstraction.
384.
by ukuina 3y ago
If GPT4 is 20x the price of GPT3.5, but it only takes 10x GPT3.5 runs to get similar quality of response (and likely faster), you'll still come out ahead.
385.
by ukuina 3y ago
I appreciate when teams provide such a detailed write-up of their tech stacks. The callouts to bad choices and recovery (e.g., Nomad) is also refreshing to see. Thank you for the share!
386.
by ukuina 3y ago
"You, too, can be a thousand-aire!"
387.
by ukuina 3y ago
Congrats! Looks great, and I can't wait to try it. Do you support AzureOpenAI with custom endpoints? Are any special settings necessary to disable telemetry or non-core network requests?
388.
by ukuina 3y ago
This makes it difficult to justify a production deployment: > The core of last_layer is deliberately kept closed-source for several reasons. Foremost among these is the concern over reverse engineering. By limiting access to the inner wo
389.
by ukuina 3y ago
Careful, Greenshot phones home. Verified with a firewall restriction.
390.
Cleft: Turn Voice Memos into Shared Notes with On-Device Transcription(cleftnotes.com)
5 points by ukuina 3y ago|1 comments

More ›