I just did a captcha the other day that asked the user to select which items can fit inside the sample item (which was a handbag). You'd think that a multimodal deep learning model could figure out what objects fit inside other objects if it's going to cure cancer or whatever, but no I'm assuming that it needs to be taught explicitly.
There's a fun experiment with toddlers where they re-enter a room but the car they just sat in was replaced by a tiny version: They will try to get into the car even though only their foot fits in.
So size/scale is not as easy a concept to model in our minds as we might assume.
Do you cache data on the client and write buffer it on the way back to the server? I made a key value store as a hobby project, that also kind of streamed the data and it was convenient to do that .
Hehe :) in that sense I'm not sure if I exist in the physical world at all. It's more of an in-my-head thing. I perceive the outer body as entirely separate from my actual self. Maybe that's due to my dissociative disorder, I dunno for sure.
The problem with natural keys is that nobody ever says, "My bad, I should have spelled my name the same way on this form as I did when I registered for the service, I promise not to do it again." Instead they say, "No matter how I spell it, you should have still known it was my name!"
that's nice, but the US digital asset stockpile is not the strategic bitcoin reserve. The sentence that controls acquisition in the strategic bitcoin reserve is the one before it:
'The Secretary of the Treasury and the Secretary of Commerce shall develop strategies for acquiring additional Government BTC provided that such strategies are budget neutral and do not impose incremental costs on United States taxpayers.'
I wonder what budget neutral means. Does it mean that they have to find spending cuts in other areas in order to acquire additional BTC?
One would think that the difficulty of making a company profitable when it's training larger and larger LLMs, combined with the diminishing returns and model collapse phenomenon, would make it so that companies don't wish to stop training larger models. I assume that they continue training larger models because whichever company stops training larger models would fall behind in the race to win new rounds of funding, but if that is the case what is the ultimate valuation that these companies are trying to achieve being valued in the biliions already?
Diminishing returns means that the user gets less marginal benefit with each larger model, and the model collapse phenomenon means that models trained on new training data might be less good than older models. Have straightforward mitigations been put in place such as filtering out from the training data forums where users like to share AI generated content?
Yes routers learn details about the destination of the packets they carry, but this is not a problem for most purposes. The author goes on to make some vague insinuation that tls implementations are insecure but if they are the solution is to fix the implementation. The Internet is designed so that all types of implementations of networking software can communicate on equal terms which means that it is already an open protocol. L take
Typically it involves some entity which can use traffic-data to strip away the privacy people think they have.
A simple example might be to determine the true identity of an HN poster: Even with the content encrypted, it's harder to hide the fact whenever Account X makes a post, Person Y is always sending packets to HN.
reply