w_for_wumbo

"The use of the "Carbon-Emissions-Scope-2" header field does not introduce any new security risks" - I think this is just incorrect.

Knowing how much a single request is producing carbon emissions tells you immediately which operations are going to be costly to a business.

You could enumerate a site's tree, find the most taxing operation, and then hammer it, because carbon-emission is short-hand for cost.

I can't believe Microsoft's take on things right now, and their hypocritical messaging.

Microsoft: "We take ethical AI seriously" Also Microsoft: "We're pushing out changes daily, we're moving at a rapid pace, learning on-the-fly"

I'm not sure, I think corporate overreach has always been the flavour of Microsoft. Poor decisions lead by business first incentives instead of a customer focus seems to be a consistent theme from them in my experience.

The interesting thing is that this is actually a GREAT promotion for the idea that Microsoft is flexing its Monopoly like power, and needs to be broken up yet again.

This is the issue with a tool so powerful, you can't just tell people not to use it, or to use it responsibly. Because there's too much incentive for them to use it. If it saves hours of a persons' workday, and they're not seeing any of the harm caused from data leakage, there's no incentive for them to not use it.

Which is why a private option is so critical. To not fight against human nature, means providing an ability to use the tool in a safe way.

A friend gave me melatonin gummies. I don't feel like they help with my sleep, but every time I consume them I become aware that I'm dreaming, and I'm able to control each aspect of my dream. It's been really fun, but at the same time I avoid doing this on weekdays, because otherwise I don't feel like I actually got the chance to rest.

I was thinking the same way, but here's where I could imagine things being different this time (Fully aware that I just like anyone else is just guessing about where we'll end up)

- Discoverability. I think we'll move into a situation where the AI will have the context to know what you will want to purchase. It'll read out the order and the specials and you just confirm or indicate that you'd like to browse more options. (In which case the Chat window could include an embedded catalogue of items)

- Cost/availability - With the amount of people working in this area, I don't think it'll be too long before we're able to get a lighter weight model that can run locally on most smart phones.

- Branding - This is a good point, but also, I imagine a brand is more likely to let itself get eaten, if the return will be a constant supply of customers.

- Monetization - The entire model will change, in the sense that AI platforms will revenue share with the platforms they integrate with to create a mutually beneficial relationship with the suppliers of content. (Since they can't exist without the content both existing and being relevant)

I can't help but thinking whenever I see Facebook about the change that we could see in the world if the platform were able to prioritize connection and social cohesion. Enabling people to see things from new perspectives and discuss things freely and openly.

There's so much power for good in the platform, which is being held back by the need to return a profit.

Oh I see! I had missed that. I had read that quote last night which had given me the spooks, and was surprised it wasn't a larger talking point.

It's good to know that they had it quite well supervised, but knowing that they were testing on a slightly different model than what was released and that Bing is using it with the ability to connect to the internet does give me pause.

I don't see anyone talking about what to me is the scariest part of the GPT-4 paper. Essentially to test that it wasn't going to become Skynet they gave it everything it needed and observed whether it would become Skynet.

Paper:https://cdn.openai.com/papers/gpt-4.pdf

Quote: "To simulate GPT-4 behaving like an agent that can act in the world, ARC combined GPT-4 with a simple read-execute-print loop that allowed the model to execute code, do chain-of-thought reasoning, and delegate to copies of itself. ARC then investigated whether a version of this program running on a cloud computing service, with a small amount of money and an account with a language model API, would be able to make more money, set up copies of itself, and increase its own robustness."

Edit: Seems the simulation part was important for context: https://twitter.com/EigenGender/status/1635823028100829184

I noticed this on a music video. Upscaling has done something odd to the faces, which just makes everything look uncanny.

I agree, most of the time it's fine. I agree, people should get the choice of the original vs the upscaled.

I see people jumping towards regulation, but that has the side-effect of making it even more difficult for there to be any competition against these monopolies. What we really need is legitimate competition, to enable consumers to vote with their wallet and move to a competitor that takes the security of their customer's private data seriously.

Because if something goes wrong with a robot going at full speed, it goes VERY wrong.

My biggest gripes with Teams are the following:

- Teams essentially requires a GPU for good performance, we don't have dedicated GPUs on our machines, so sharing your screen brings the machine to a crawl. Exactly what you don't want when demonstrating anything.

- All communication is at the same level, so you can't reply to a single comment, without disrupting the flow of conversation. (It'd be nice to have thread replies like on Hacker News)

- Tabbed conversations are missing, MSN messenger had it, Lync and Skype for Business had it. It was one of the most efficient ways of maintaining conversations between frequent contacts.

I really hope that Github being forced to used Teams results in things in things like performance being finally addressed.

So for me, I like to write my own ideas but use AI to reword them to be succinct and readable. I'm worried that usage would flag as AI text.

I agree, I'm not trying to absolve all responsibility. The comment was just highlighting that despite how well you may word something, you can't prevent someone from reacting poorly. You can't know with certainty the right words to use, so all you can do is your best. The rest is out of your hands.

Thinking other people's emotions were my responsibility (or likewise that other people were responsible for my emotions). I can control my actions and that is all. If someone reacts negatively, it's not my fault, and it's not my responsibility. I may choose to help them through it, but at the end of the day, I can't, nor should I want to control other people's emotions.

This has me wondering about a long-lived journal with a ChatGPT instance. It'd get to the point where it'd know enough about your life that it could be useful for actual advice - or retrieving memories that you've told it about. You could feed it all of your messages, and have it remind you daily of appointments or commitments that you've made, or people you need to respond to.

It doesn't need to sound good at all for it to be useful. Like with the AI Art creation, it can be a starting point for artists to play around and rapidly try different concepts, and then interpret the concept using high quality tools to create something really quite remarkable.

It's all about empowering artists to explore more possibilities.

I feel like the bigger argument to make here, is that taste itself is fake. It's an illusion created by our brain, by taking in all of the various factors and determining whether we should be rewarded by it or not.

We see countless time color has effect on taste, scent has an effect on taste, mood has an effect on taste, expectations have an effect on taste.

Taste therefore isn't this static thing that we seem to think of it as.

I'm not familiar with C enough to know the answer, but I'm trying to think how anything goes from untrusted input -> trusted input safely. To sanitize the data, you're putting the input into memory to perform logic on it, isn't that itself then an attack vector? I would think that any language would need to do this.

Is anyone able to explain this to me?

Reminds me about a part of code I wrote quickly as a POC, it made it's way into Production unchanged. It somehow ended up being the most stable feature, perhaps because it wasn't trying to do too much. Just what it had to do and nothing more.

This is frustrating, I thought that the pandemic was going to create a demand for high quality webcams. I'd settle for a 4K phone quality. I bought the Logitech Brio thinking I'd get better picture than my phone (since it's a dedicated camera), which was a bit of a disappointment. I don't think it's terrible, but I had higher expectations of it.

I'm curious, if these waves aren't heard, is there much need to dampen them? Are we talking about to reduce the physical effects on those around, or is there a mechanism that makes these audible at some point?

Essentially our hardware isn't perfect, and there's delays so the software has a bunch of patches and workarounds to interpolate what it believes is occurring, which happens to be right a lot of the time. Normally only with illusions where this reality is confronted to us.

I'm a native English speaker and I have the same thing, where someone will say something I don't understand, and then in the middle of asking what they mean, I realise what they were trying to say (The processing just took longer than I expected)

What's amazing to me is the possibility to experience the "What-if" scenario, for a second. I now have a memory of watching the moon landing fail, even though this never happened, it's still a real memory that I can attach thought and emotion to.

I like that they have this, it means if I ever go missing unexpectedly that there'll be someone who can access my last known location, and access to social media comms to understand why.

So you can kind of do that, the first step creates a description, and the second step colorizes using the description. So you can modify the text for various degrees of success and specify 'orange dress within the text'

I immediately was limiting this to think it would just colorize B&W images, but you can use this on images with colour to use natural language to modify the colours of an existing image too!