> avoid onions and carrots because it causes death
onions and carrots do not experience suffering. even if they did, most plants are grown for non-human animal feed. so the best way to reduce "plant suffering" would be to stop exploiting non-human animals.
This is so cool! It's a little unclear from the website, is there hardware I could buy today to physically remove carbon dioxide? What can I actually do?
(I recommend removing 1 kg just for the experience, it's $1)
In terms of buying the hardware yourself, the low end is still tens of thousands of dollars. If you happen to own a farm, a pyrolysis machine can turn your farm waste into solid carbon that's locked up for hundreds of years. And people will pay you to do it too.
And I love your question, I really want individuals to be able to take clear action on this. We've built tools for that, clearly we're not there yet (higher CO2 emissions than ever last year) but not yet "plug this thing in at home".
The other thing that comes to mind is we have a online course where thousands of people have got together and learned about how carbon dioxide removal works: https://bootup.airminers.com/
It's book club style, so you're matched up with a bunch of people and you get to talk through the topics together. Let me know let you think about all this!
I don't think we'll ever reduce our CO2 emissions - I've done a couple of decades of sustainability organizing (and professionally, I ran Amazon's initial carbon accounting program) and the best I've seen anyone model is "slightly less growth in CO2 emissions." I suspect we're stuck with sucking it out, and resilience efforts.
I'll check out the online course, thank you. I saw there are a few startups offering capture at an office building scale, and I'll come back with better questions once I learn more!
GENEVA – Israel has committed genocide against Palestinians in the Gaza Strip, the UN Independent International Commission of Inquiry on the Occupied Palestinian Territory, including East Jerusalem, and Israel said in a new report today. The Commission urges Israel and all States to fulfil their legal obligations under international law to end the genocide and punish those responsible for it.
The UN, like most western organizations, is biased in favor of Israel. Meaning, this reporting is what favorable coverage looks like. You should read unfavorable coverage, it's much much worse. It took the UN a long time to acknowledge Israel even did anything wrong.
The UN is a political tool absolutely - and it's one that's biased in favor of Israel.
Look, you have to understand that you and I are looking at this from a western lense. The west is EXTREMELY pro-israel. So your Overton window isnt just shifted a little buddy. It's fucking whiplashed alllllll the way to one side. So yes, from your perspective or mine the UN might appear as not pro Israel.
But um, they are, and quite a bit too. Again, if you want to see actual real anti-israel reporting, read stuff from Iran. You will surely notice it sounds very, very different from how the UN portrays Israel.
It's turtles all the way down. If we had a way to perfectly prevent people from doing undesirable things, we wouldn't need courthouses to begin with. The system doesn't need to be perfect, it just needs to be good enough that reliably circumventing it isn't worth the effort.
It sure changes the incentives though. It’s much less attractive to leak recordings as a PR move—or realize any benefit that cranky humorless judges can trace back to the recording—if that, in and of itself, constitutes a whole new crime (and effectively confessing to it too).
An operator at load capacity can either refuse requests, or move the knobs (quantization, thinking time) so requests process faster. Both of those things make customers unhappy, but only one is obvious.
>And according to Google, they always delete data if requested.
However, the request form is on display in the bottom of a locked filing cabinet stuck in a disused lavatory with a sign on the door saying ‘Beware of the Leopard'.
LLMs sample the next token from a conditional probability distribution, the hope is that dumb sequences are less probable but they will just happen naturally.
I wouldn't doubt that these companies would deliberately degrade performance to manage load, but it's also true that humans are notoriously terrible at identifying random distributions, even with something as simple as a coin flip. It's very possible that what you view as degradation is just "bad RNG".
Thats what is called an "overly specific denial". It sounds more palatable if you say "we deployed a newly quantized model of Opus and here are cherry picked benchmarks to show its the same", and even that they don't announce publicly.
Personally, I'd rather get queued up on a long wait time I mean not ridiculously long but I am ok waiting five minutes to get correct it at least more correct responses.
I've seen some issues with garbage tokens (seemed to come from a completely different session, mentioned code I've never seen before, repeated lines over and over) during high load, suspect anthropic have some threading bugs or race conditions in their caching/inference code that only happen during very high load
If you use the API, you pay for a specific model, yes, but even then there are "workarounds" for them, such as someone else pointed out by reducing the amount of time they let it "think".
If you use the subscriptions, the terms specifically says that beyond the caps they can limit your "model and feature usage, at our discretion".
Sure. I was separating the model - which Anthropic promises not to downgrade - and the "thinking time" - which Anthropic doesn't promise not to downgrade. It seems the latter is very likely the culprit in this case.
Old school Gemini used to do this. It was super obvious because mid day the model would go from stupid to completely brain dead. I have a screenshot of Google's FAQ on my PC from 2024-09-13 that says this (I took it to post to discord):
> How do I know which model Gemini is using in its responses?
> We believe in using the right model for the right task. We use various models at hand for specific tasks based on what we think will provide the best experience.
No, basically, the requests are processed in batches, together, and the order they're listed in matters for the results, as the grid (tiles) that the GPU is ultimately processing, are different depending on what order they entered at.
So if you want batching + determinism, you need the same batch with the same order which obviously don't work when there are N+1 clients instead of just one.
Small subtle errors that are only exposed at certain execution parts could be one. You might place things differently onto the GPU depending on how large the batch is, if you've found one way to be faster batch_size<1024, but another when batch_size>1024. As number of concurrent incoming requests goes up, you increase batch_size. Just one possibility, guess there could be a multitude of reasons, as it's really hard to reason about until you sit with the data in front of you. vLLM has had bugs with these sort of thing too, so wouldn't surprise me.
No, I'm not sure how that'd make sense. Either you're making the correct (expected) calculations, or you're getting it wrong. Depending the type of wrong or how wrong, could go from "used #2 in attention instead of #1" so "blue" instead of "Blue" or whatever, to completely incoherent text and garbled output.
I accept errors are more likely to decrease "intelligence". But I don't see how increased load, through batching, is any more likely to increase than decrease errors.
onions and carrots do not experience suffering. even if they did, most plants are grown for non-human animal feed. so the best way to reduce "plant suffering" would be to stop exploiting non-human animals.