peaslock

Maybe not a good idea to link a page that runs source code created by random people. Well CSS is very safe, but still.

Not if you ask first.

Neural nets often fail with (repetitive) gibberish output when the input is too different from the training data. This model appears to take in the entire text input at once or look ahead at the next input letters, so the unusual "bla bla" at the end can mess up outputs near the beginning.

Can DRIZZLE help to achieve higher resolution? Though with hundreds of photos this will imply a lot of work:

https://en.wikipedia.org/wiki/Drizzle_(image_processing)

Geoffrey Hinton has recently been talking about how analog and "imperfect" computing with specialized hardware/circuitry may yield much cheaper neural nets, that could easily be as large as human brains, but would only cost a few dollars and would be extremely cheap to run. Not a new idea, but it is a fairly promising outlook, I think.

https://www.zdnet.com/article/we-will-see-a-completely-new-t...

But they have highly likely internal prototypes with higher bandwidth and latency. Also, with distilled latent diffusion one can probably generate text(-images) much faster anyhow as it could produce long chunks of text at once, rather than needing recurrently feed back the new token to the inputs.

Yes, but the losses in Figure 3 increase because the larger models see fewer data to keep the FLOP budget constant, not because of overfitting. Large models do not overfit very much, so the loss of a larger model will still be better compared to a smaller model when you keep dataset size constant.

Though isn't it highly likely that core devs working at the big tech giants have access to 10x-100x faster compute, e.g. some secret TPU successor at Google?

> if you want improved performance, you still need more data

Not true. See figure 2: https://arxiv.org/pdf/2203.15556.pdf#page=5

The loss decreases with greater model size at the same compute budget (i.e. stopping sooner regarding training data). Also some rehearsal/multi-epoch training improves the forgetting rate (thereby improving performance substantially), which hasn't been taken into account by Chinchilla et al. because they train <1 epoch.

https://arxiv.org/abs/2205.12393

Not necessarily: https://arxiv.org/abs/2206.14486

Also, even with "Chinchilla laws", you still gain performance in a larger model, you just need a lot more data (if just as noisy) to reach the same level of convergence, but a larger model will have already partially converged to a superior model with the same amount data.

Yeah, continuous online learning by fine-tuning seems like an obvious way of making these models recall information from outside the perceptible context. One could also prompt the model to (recursively) summarize code and prepend this summary to each prompt, and/or enable the model to interactively query function definitions or code summaries before outputting a final answer (trained by RLHF). But any such tricks might also quickly be outcompeted by an even more general model, e.g. one that directly controls the GUI and can communicate with coworkers...

The model with the most similar name in this list is code-cushman-001 which is described as "Codex model that is a stronger, multilingual version of the Codex (12B) model in the paper".

https://crfm-models.stanford.edu/static/help.html

The next stronger Codex model is called code-davinci-001 which appears to be a fine-tuned version of the GPT-3 Davinci model which is known to have 175B parameters. The model naming is alphabetical in the order of the model size:

https://blog.eleuther.ai/gpt3-model-sizes/

See also A.2 here: https://arxiv.org/pdf/2204.00498.pdf#page=6

Amazing if this is only a 12B model. If this already increases coding productivity by up to 50% (depending on kind of work), imagine what a 1T model will be capable of! I do wonder if some programmers at FAANG are already having access to a way more powerful coding assistants, and whether they code much at all at this point, or only make high level code specifications and then fix up the automatically generated code.

Can you use a different Wi-Fi network at the same time for internet access or does Wi-Fi Direct block any other Wi-Fi access?

> Plus you don't need to be on the same network

You mean in case of the Wireless Display Adapter? For Miracast you do need to be on the same LAN, right?

Google will not disappear. They already have much larger neural nets almost ready for deployment, plus they will be able to afford even larger ones in future. And size is all that matters while the techniques are mostly trivial.

How will we justify our existence unable to contribute meaningfully to the economy?

My Firefox is now stuck with a note at the top of the screen saying something like "voxelchain.app is controlling your mouse cursor. Please press ESC to take over."

It is measurable, but not harmful to a meaningful extent. There are lots of sources of low-dose radiation in the natural human/primate/.../mammalian environment.

Speaking of which, is anyone aware of example code using the LSTM? I've been trying to get this to work, but there seems to be information missing e.g how to setup the input/output descriptors and how to manage input data: https://developer.apple.com/documentation/accelerate/bnns/us...

How long does the battery last for you browsing like that?

So 0.005% at 560 TWh total consumption.

> About third of what was used before.

How much is that as a fraction of the total electricity consumption?

How much electricity can plausibly be saved this way? And how does that compare to the energy required to reprogram and possibly renew all these signs?

While automatic data collection might become harder, one can still curate high quality datasets. An example is Tesla's FSD autopilot which is nearly entirely trained on curated data (AFAIK), as well as highly realistic 3D simulation data. Sure, it is expensive, but the expected returns are also very high.

However, there is some evidence that NNs currently are somewhat limited by the availability of high quality data,[0] however I'm not sure this is really a problem because neural nets already accomplish amazing things, so one might not need that much data to get something useful (perhaps at the expensive of more compute, but so what; e.g. analog computing might give some 1000x speedup anyhow).

[0] https://www.lesswrong.com/posts/6Fpvch8RR29qLEWNH/chinchilla...

One piece of evidence is that authorities pushing hard for surveillance do not actually care about removing images they find, even though removal requests are easy to file and are obeyed by hosters.[0]

More evidence is that this sort of crime is being exaggerated in its severity. E.g. this study found the risk of adulthood psychological trauma is minimal when relying on objective evidence of abuse.[1] It is filthy and bad, but the total damage seems hardly enough to justify abolishing privacy.

There is always going to be a tradeoff between safety and freedom.

[0]: https://news.ycombinator.com/item?id=31353026

[1]: https://www.nature.com/articles/s41562-020-0880-3

Power concentration is inevitable regardless of the system because in order to enforce decentralization you would need to, well, concentrate power.

Our best bet is to shame those using products of unethical companies.

Perhaps they induce immunity to novel viruses by transferring them in small doses between very different animals. Just guessing though.