tikkun

8 ms·

31.
by tikkun 2y ago
[Submitted on 23 May 2024]
32.
by tikkun 2y ago
It's frustrating how many disagreements come down to framings rather than actual substance. His framing of intelligence is one thing. The people who disagree with him are framing intelligence a different way. End of story. I wish that
33.
by tikkun 2y ago
None of the broad ai agent tools seem to work. Perplexity is my favorite narrow ai agent tool, kind of like an ai agent as an alternative to google search
34.
by tikkun 2y ago
I polled my friends and there were two major themes: 1) Claude/ChatGPT (copy and pasting back and forth) 2) Cursor
35.
by tikkun 2y ago
When I Say No I Feel Guilty. Never Split The Difference. The Gift of Fear. Those 3.
36.
Don't Jade Me Bro(gpus.llm-utils.org)
5 points by tikkun 2y ago|0 comments
37.
Comet Tsuchinshan-Atlas(skyandtelescope.org)
4 points by tikkun 2y ago|0 comments
38.
by tikkun 2y ago
FAQs I found: Is it available for use now? Nope When will it be available for use? On FB, IG and WhatsApp in 2025 Will it be open sourced? Maybe What are they doing before releasing it? Working with filmmakers, improving video quality, redu
39.
by tikkun 2y ago
Related, thanks to Ethan for mentioning this: https://news.ycombinator.com/item?id=41644315 Also related TrueBill, Rocket Money (thanks to Ron for mentioning these 2), and DoNotPay, but they don't quite solve this. May
40.
2-Clicks to Subscribe, 20-Clicks to Cancel, We Need an AI Subscription Reaper(gpus.llm-utils.org)
3 points by tikkun 2y ago|2 comments
41.
OpenAI Spinning Up in Deep RL Workshop (2019) [video](youtube.com)
1 points by tikkun 2y ago|0 comments
42.
Ask HN: Do you find LLMs to be judgmental and have you fixed with instructions?
1 points by tikkun 2y ago|0 comments
43.
by tikkun 2y ago
Makes me think of: some of Taleb's ideas about just-in-time manufacturing (no slack eg covid supply shocks) https://www.lesswrong.com/posts/yLLkWMDbC9ZNKbjDG/slack Also, can't recall it but a long time a
44.
by tikkun 2y ago
See here https://news.ycombinator.com/item?id=36195527
45.
Ask HN: How are you planning around AI uncertainty?
3 points by tikkun 2y ago|4 comments
46.
by tikkun 2y ago
> We need to act wisely but with conviction. Reminds me of these quotes from Sam on this podcast episode ( https://www.youtube.com/watch?v=KfuVSg-VJxE ) * "Prudence without fear" (Sam referencing another quote) *
47.
by tikkun 2y ago
> If we want to put AI into the hands of as many people as possible, we need to drive down the cost of compute and make it abundant (which requires lots of energy and chips). If we don’t build enough infrastructure, AI will be a very lim
48.
by tikkun 2y ago
The future will be abundant, because deep learning works. To achieve that, we need to be calm, but cautious. And, we need to fund infra (chips and power) so that AGI isn't limited to the ultra-wealthy. My take: * Foom/doom isn&#x2
49.
by tikkun 2y ago
I'll note that both photoshop and changing the timestamp of images (mentioned below) are only easy for a very small percentage of the population. It'd likely be different if >30% of people could easily do these things.
50.
by tikkun 2y ago
Is this Jim Neilsen related to Jakob Neilsen (famous UI guy)?
51.
by tikkun 2y ago
Overview: https://storm-project.stanford.edu/research/storm/
52.
by tikkun 2y ago
I'm optimistic that hallucination rates will go down quite a bit again with the next gen of models (gpt5 / claude 4 / gemini 2 / llama 4). I've noticed that the hallucination rate of newer more SOTA models is much l
53.
by tikkun 2y ago
> I can't help myself but think that some fundamentals changed over the past 2-3 years in the software space. LLMs enabled many new companies & created demand for more products/startups.
54.
Ask HN: Which of these product ideas would you bet on?
2 points by tikkun 2y ago|1 comments
55.
by tikkun 2y ago
Related note: It'd be great if cities were more child friendly. A big challenge is cars, the 4000+ lb fast moving metal boxes... It'd be great if there were more car-minimal areas in cities, enclosed areas with parks and restauran
56.
by tikkun 2y ago
tough to answer. depends on the llm, on the api provider or hosting method, on the prompt but you could get a simple estimate by running a CoT prompt, copying the output tokens, and then using an online token counter to count the length &am
57.
by tikkun 2y ago
mostly they're going straight to code, designing in code it'd be cool to have generative ux tools for wireframing though, if that's what you're getting at
58.
by tikkun 2y ago
"Everyone just says scaling hypothesis. Everyone neglects to ask, what are we scaling?" [Sutskever] said. Any guesses?
59.
Policy Ideas for a Healthier America(justinmares.substack.com)
2 points by tikkun 2y ago|0 comments
60.
by tikkun 2y ago
Sonos owner; I strongly dislike it. I disliked it on the old app, too. Will get Apple homepods next.

More ›