chaoz_

8 ms·

31.
Mechanical Turk: 84 Years of Chess and Deception(en.wikipedia.org)
2 points by chaoz_ 2y ago|0 comments
32.
Show HN: CaptureFlow – Provide LLM with debugger-level context of your app(github.com)
3 points by chaoz_ 3y ago|0 comments
33.
by chaoz_ 3y ago
We have no good benchmark to estimate the bugfixing ability, it was mostly zero-short "in this case it works" example.
34.
Show HN: CaptureFlow – LLM codegen/bugfix powered by live application context(github.com)
6 points by chaoz_ 3y ago|3 comments
35.
by chaoz_ 3y ago
I've been using pyenv for several years now, but for some reason, the basic commands and overall integration don't feel as smooth as Node's nvm package. I wonder if that's because Python setup is technically harder than
36.
by chaoz_ 3y ago
I think the idea is that after certain tech piece is getting regulated, giants will use this as opportunity to push for some replacement which they will get to influence. https://www.theverge.com/2023/7/20/238
37.
AWS Card Clash [video](youtube.com)
1 points by chaoz_ 3y ago|0 comments
38.
by chaoz_ 3y ago
Love the idea, however, it's not clear whether I will get access to a large collection of components for building such workflows or what is currently possible? Would nice to get this info before proceeding with auth.
39.
Asynchronous Local-SGD Training for Language Modeling(arxiv.org)
1 points by chaoz_ 3y ago|0 comments
40.
by chaoz_ 3y ago
Sounds like it's still an acceptable cost of doing business.
41.
GPT-4: Consistent Karma Generator for Reddit(nikitakutz.substack.com)
4 points by chaoz_ 3y ago|2 comments
42.
by chaoz_ 3y ago
Hi! I was wondering if GPT4 can easily assimilate to diverse set of subreddits, and the results confirm it. Really hope Reddit won't get spammed with product advertisement agents, I still find it better than google for queries that got
43.
by chaoz_ 3y ago
Hi everyone! Have you wondered if AI can assimilate into Reddit community? Conducted a mini-experiment using LLM-based characters to navigate through subreddits, and the results were pretty interesting.
44.
by chaoz_ 3y ago
Will output quality of an average excel user go down, if the job duties will evolve into verifying LLM output and not performing low-level tasks over and over? Laziness is a safe bet, so we will see :)
45.
by chaoz_ 3y ago
With the current state, most powerful LLMs are also limited in the amount of topics they are allowed to discuss. I bet you could easily differentiate between player and AI by asking opinions on some controversial topic.
46.
by chaoz_ 3y ago
True... When I measured something similar in a large python app, the biggest chunk of time went into python object serialization/deserialization.
47.
by chaoz_ 3y ago
Researching other options will take some time + the deliverability/spam-detection will only get worse. You can check e-mail price of "@gmail" on darkweb and compare it against other providers. For them it's a "thres
48.
by chaoz_ 3y ago
You can always check Qt TextEdit source code ;)
49.
by chaoz_ 3y ago
right, or giant corps will treat such startups as marketplace app creators for their LLM API and kinda "outsource" the innovation, while providing the platform and enjoying API income?
50.
by chaoz_ 3y ago
Will definitely use that for my next dopamine detox cycle.
51.
by chaoz_ 3y ago
This does not include ChatGPT traffic tho :)
52.
by chaoz_ 3y ago
Damn I saw too many fun VK apps, which ones have you authored? :D
53.
by chaoz_ 3y ago
https://patents.google.com/patent/US10452978B2/en GPT models are based on transformer, but architecture is different from what's patented. Not a lawyer, but can you really patent certain network architecture?
54.
by chaoz_ 4y ago
True, "amount of hallucination" (very confident, but factually wrong) is probably something they can decrease in the next versions tho. I also would not trust it with anything important, but there can be good applications for some
55.
by chaoz_ 4y ago
Rent is not cheap, but if you are okay with 1 hour commute you can live in the Hague, Leiden or Utrecht. Less tourists also. Btw, there are large companies in AMS (Booking, Amazon, Meta, Uber) and some financial shops. It's not compara
56.
by chaoz_ 4y ago
Yes, it is sophisticated. Phishing emails are opened more often then regular mail (assuming it’s in your inbox) and click-rate is pretty high. You might be good at cracking phishing simulation at your workplace, but many people won’t know i
57.
by chaoz_ 4y ago
or it's just a good bet to maximize chances. It's not like they invested a large portion of their assets. edit: last sentence sounds kinda weird given recent layoffs round
58.
by chaoz_ 4y ago
"We're all so comprehensively pwned that if an email address isn't pwned, there's a good chance it doesn't belong to a real human." https://www.troyhunt.com/pwned-or-bot/
59.
by chaoz_ 4y ago
I found Andrej Karpathy's explanation of transformers to be very insightful: https://www.youtube.com/watch?v=9uw3F6rndnA
60.
by chaoz_ 4y ago
I really want to agree, but those are different tasks. Apple and Amazon attempt to focus on monetizable parts of NLP flows meaning optimization for narrow use cases (play music, order me a book, etc) and OpenAI can share an impressive chat

More ›