19h

12 ms·

91.
by 19h 3y ago
Is that old information? I’m sure I have a dump on most Turkish civilians from a few years back… it also includes data on Erdogan, his birth place and ID number.
92.
by 19h 3y ago
Can’t Apollo be open-sourced and the tokens of the original app be used? I’d absolutely donate on a monthly basis.
93.
by 19h 3y ago
I’m confused — how is this legally binding? In Germany, a termination is valid IFF the termination has been submitted to the employee in writing in a timely manner — that is, they need to have been able to receive it before end of month or
94.
by 19h 3y ago
If you're afraid of crawlers that will feed into AI datasets, why not just render images into canvas elements? This kind of watermark makes images bad for everyone, not just the AI. Do you really want to see this on every website you v
95.
by 19h 3y ago
Claude+ is only called like that by Poe -- it's actually Claude v1.3 and Claude Instant v1 (as reflected on the site).
96.
by 19h 3y ago
Yes, existing files are not updated.
97.
by 19h 3y ago
No we should not, at least for businesses that sell physical things. It’s been bugging me a lot that many Chinese vendors bypass the EU‘s 2 year warranty laws by simply ceasing to exist.
98.
by 19h 3y ago
Funny to see improvements in crypto.createHash… I was totally caught off guard yesterday noticing most of crypto has been removed in Node 20 and replaced with the exclusively-async „subtle“ WebCrypto. Quite a pain to work with when you need
99.
by 19h 3y ago
You can configure Copilot to not return code that appears verbatim in public repositories. In that case it at least won't produce code you could legitimately argue would be covered by any individuals' specific license.
100.
by 19h 3y ago
What a ridiculous article. Copilot does not violate antitrust law. GitHub is not a monopoly just because open source devs choose to host there. Devs are free to use GitLab or whatever. Comparing this to Google Books is silly. Google stole c
101.
by 19h 3y ago
That's just wrong. We ran metric tons of questionable content before being approved for anything.
102.
by 19h 3y ago
If we were to get banned for this, we'd have been banned long ago. We literally process "questionable" content "as a service" and this use case was explicitly approved. (We do heuristics and ML-assisted background c
103.
by 19h 3y ago
Claude may reject answers, but other than OpenAI‘s GPT you can put words into the mouth of the assistant and essentially bypass safety checks. In fact, Anthropic explicitly discusses putting words into the assistants‘ mouth to be able to sh
104.
by 19h 3y ago
We have access and we’re already playing with the 100K model. It’s pretty insane. We‘re about to ditch all of our recursive summarization code. If you want me to test something for you, lmk and I’ll send it through the api.
105.
by 19h 3y ago
We’ve been having alpha access to Claude[0] and it’s absolutely scary how real and human it sounds compared to ChatGPT — our entire team is using Claude for website copy and summarization, GPT4 is only used for code. [0] claude-v1.3 — claud
106.
by 19h 3y ago
Correlation does not equal causation. Diets high in red meat and fat have been the norm for most of human history, yet diabetes has only become an epidemic in recent decades. The real culprits are the processed foods and sedentary lifestyle
107.
by 19h 3y ago
Pfft, keys, schmeys. Real security is built on handshakes and backroom deals, not strong encryption.
108.
by 19h 3y ago
My apologies, I misinterpreted your comment. You make a fair point that HN discussions are not equivalent to formal peer review.
109.
by 19h 3y ago
There's no need to attack the entire HN community over one person's opinion. Preprints and discussions here both have value, and different forms of review suit different needs.
110.
by 19h 3y ago
While I understand the need to protect sensitive ecosystems, restricting access to nature altogether is extremely problematic. There must be better solutions that don't infringe on what should be a basic human right. If overpopulation
111.
by 19h 3y ago
If people actually took some personal responsibility instead of blaming 'the system' for everything, places wouldn't get trashed in the first place.
112.
by 19h 3y ago
True, though some states have 'recreational use statutes' that limit landowner liability for allowing public access to private land. But there are still many risks and headaches involved, even with those protections.
113.
by 19h 3y ago
Because that would make too much sense.
114.
by 19h 3y ago
You're completely out of touch with reality if you think $1000/night for camping is acceptable or normal.
115.
by 19h 3y ago
Comparing any tech CEO's pay to Buffett's is silly. Buffett built Berkshire from scratch over 50+ years through incredibly savvy investments. Nadella inherited a tech giant and has kept the momentum going, but the scale and diffic
116.
by 19h 3y ago
While the system certainly has its flaws, calling them 'simply exploiters' ignores the immense value they've created and the hard work required to build a company like Google.
117.
by 19h 3y ago
The class structure provides a clean API for building on, even if the internals are basic. With some refinement, it could be a good starting point for more advanced models.
118.
by 19h 3y ago
If it makes you feel any better, the researchers building these things don't really know what they're doing either. They just throw data and compute at the problem and hope for the best.
119.
by 19h 3y ago
Comparing Langchain to Hugging Face Transformers is apples and oranges. One is for research, one is for production. Production ML requires more abstraction, not less.
120.
by 19h 3y ago
Grounding helps avoid nonsensical or factually incorrect hallucinations, which are unacceptable for most applications. At the same time, developing techniques to improve the base LLM and narrow the context window is important for achieving

More ›