bjourne

36 ms·

451.
by bjourne 1y ago
IMnsHO, nope! People trust the bot way more than their own experience.
452.
by bjourne 1y ago
Does having your personal website even matter when the agents of censorship can just request that search engines delist your urls? Or pay for tons of ads so that your site's ranking drops to the second or third page for whatever keywor
453.
Why did Wikipedia editor block edits to the 'Gaza genocide' page?(aljazeera.com)
2 points by bjourne 1y ago|0 comments
454.
by bjourne 1y ago
> 1. Conflates consciousness with "thinking" - LLMs may process information effectively without being conscious, but the article treats these as the same phenomenon There is NO WAY you can define "consciousness" in su
455.
by bjourne 1y ago
"have made this type of content relatively easy to churn out on demand": It doesn't say the papers are LLM-generated .
456.
by bjourne 1y ago
Not what the announcement says. And if you're so sure it's possible, show us how it's done.
457.
by bjourne 1y ago
If you believe that, can you demonstrate how to generate a position or review paper using an LLM?
458.
by bjourne 1y ago
Would the best version of yourself write such a manual?
459.
by bjourne 1y ago
So Microsoft is nowadays the American "do no evil" tech giant. How the times have changed!
460.
by bjourne 1y ago
Of course there is red tape. EU funding comes from taxpayer money and we want it to be spent wisely. The red tape is precisely to prevent it from being funneled to pals. EU has funded quite a few free software projects so it's not like
461.
by bjourne 1y ago
A spectrogram is lossy and not a one-to-one mapping of the waveform. Riffusion is, afaik, limited to five-second-clips. For these, structure and coherence over time isn't important and the data is strongly spatially correlated. E.g., a
462.
by bjourne 1y ago
You are talking about piano roll notation, I think. While it's 2d data, it's not quite the same as actual image data. E.g., 2d conv and pooling operations are useless for music. The patterns and dependencies are too subtle to be c
463.
by bjourne 1y ago
> My advice would be 1) to not phrase decisions as suggestions, I was about to write the same thing. TS, if you feel strongly about the need for a post-mortem analysis then it is: "We WILL do a post-mortem analysis, and you WILL ana
464.
by bjourne 1y ago
How is he disrespecting you? "Respect" means so many different things to so many people that it is difficult to say what it is about his behavior you find intolerable. Maybe the answer is for you to grow a thicker skin because you
465.
by bjourne 1y ago
> Monarch lets you program distributed systems the way you’d program a single machine, hiding the complexity of distributed computing: There are some infamous tech based on the "hiding" paradigm. PHP comes to mind. By hiding ho
466.
by bjourne 1y ago
If people can comprehend your code they can point out flaws in it that invalidate your experiments. But be a good researcher and don't think like that. :)
467.
by bjourne 1y ago
You're probably right. I trusted Google's summary: "Rupert Murdoch's News Corp owns the British newspaper The Times and its sister paper, The Sunday Times. Both are published by Times Newspapers Ltd, a subsidiary of News
468.
by bjourne 1y ago
British tabloid owned by Rupert Murdoch.
469.
by bjourne 1y ago
This is obvious marketing spam. Not hn worthy imo.
470.
by bjourne 1y ago
These blog posts might interest you. Note the dates: https://scottlocklin.wordpress.com/2010/10/04/spotting-vapor... https://scottlocklin.wordpress.com/2019/01/15/quantum-comput
471.
by bjourne 1y ago
Well, that an autoregressive model has parameters does not mean it is not autoregressive. LLMs are not Markovian.
472.
by bjourne 1y ago
If so they are wrong. :) Autoregressive just means that the probability of the next token is just a function of the already seen/emitted tokens. Any "ideas that may exist" are entirely embedded in this sequence.
473.
by bjourne 1y ago
That is precisely what autoregressive means. Perhaps you meant to write that modern LLMs are not strictly autoregressive?
474.
by bjourne 1y ago
1. Yes 2. No
475.
by bjourne 1y ago
I'm not a giant like Schmidhuber so I might be wrong, but imo there are at least two features that set residual connections and LSTMs apart: 1. In LSTMs skip connections help propagate gradients backwards through time . In ResNets, sk
476.
by bjourne 1y ago
No, I meant exactly what you wrote. You say that since physical strength of men and women differ their cognitive abilities should differ too. The physical strength of muscular and skinny men differ too, hence, according to your argument, th
477.
by bjourne 1y ago
Would you assume the same thing about men? That a muscular man has different "societal expectations" etc. than a skinny man?
478.
by bjourne 1y ago
Environment variables are not more secure than command line parameters! It's such a common misconception that because environment variables are "not seen" they can serve as a secure channel. They emphatically cannot.
479.
by bjourne 1y ago
No, the private keys are not encrypted. If there were, where would you store the encryption key for decrypting them?
480.
by bjourne 1y ago
"MY_APP_SECRET_KEY=bla myapp" hardly accomplishes anything substantial over "myapp --secret-key=blah" It's just a less robust and less well-supported command-line interface. It's not supported by most gui launc

More ›