Honestly if you can get yourself a video card with at least 12gb of RAM and I highly recommend running open source models locally.<p>It doesn’t have to be the latest fancy gaming card - I run an old 2016 Tesla P100 ($200AUD~) which is fine for 16B q4 models.<p>They’re more configurable, there’s no subscription model, you contribute to them and meaningful ways, there is a privacy concern if you’re passing your own data and it’s perhaps strangely a more fulfilling experience.<p>I’ve got mine hooked up to Home Assistant so I can automate things with Siri and shortcuts, I’m using it finding patterns and code and several decades of digital notes.<p>If you’re not sure where to start, you can try some models out locally on your machine, using apps like <a href="https://lmstudio.ai" rel="nofollow noreferrer">https://lmstudio.ai</a> or GPT4All with almost no ML knowledge. Then if you want to level up and customise things and run them on the server you can use localAI, Text generation web UI etc which provide openAI compatible APIs.<p>A good 13B model with decent prompting feels pretty on par with chatGPT (3.5) for general tasks, but can be configured to be quite good at very specific things.<p><a href="https://github.com/oobabooga/text-generation-webui">https://github.com/oobabooga/text-generation-webui</a><p><a href="https://github.com/go-skynet/LocalAI">https://github.com/go-skynet/LocalAI</a><p><a href="https://github.com/gitmylo/audio-webui">https://github.com/gitmylo/audio-webui</a>
> Don’t miss out on the latest developments in business and finance. Let us notify you as they happen.<p><i>press “I'm in”, just to mess around</i><p>> Push Notifications are available for only chrome browser right now. You will be redirected to home page in few seconds.<p>This is some advanced level of bullshit.
*Could* cost *up to* $700k/day *according to an outsider analyst*<p>Also, nowhere did the original source say that this would lead to bankruptcy. This article is bs.
This sounds like BS. Lets assume investor money covered the R&D costs and assume all other numbers are correct. 10Bn dollars would be enough for 31 years of 700kUSD spend every day...<p>This ignores the fact that quite a lot of people pay for ChatGPT pro. That might not cover the costs but I'm sure their burn rate won't expire soon. 10Bn is a LOT of money.
Is it that expensive, unoptimized due to very early or the costs are also include the training for new model?<p>Wonder what'll be the minimum future cost when it's already optimized and used for self hosted commercial purpose.
First off it would never just go bankrupt just for the reason it would be one hell of an acquihire. let alone someone would pay for the ChatGPT tech.<p>I doubt they have alot of actual debt.<p>they have paid users. People pay to use the API.<p>I think ChatGPT was an accident and was rally intended to be a "demo"<p>If chatGPT can't raise more money then no one can. You'd be crazy not to make a 1-5% of your wealth bet on GPT5.<p>if they went public today I would invest in them before anything else. I would argue a big part of microsofts recent bump is stock market access to chatgpt.
The world will not stand for losing the newly gained capabilities of ChatGPT and the like. The fact that it costs $700k a day to run shows that there is still a very high cost of entering the space, and so the incumbent players have the advantage.<p>There are serious military advantages to having access to a machine learning model this advanced. Not only propaganda, but actual war planning. The only thing that's going to close this box is world war, or a catastrophic natural event.
“If the company has 100 million users as it has claimed, and converted 1 percent of them to paying customers, that’s good enough to put them on a $200 million plus run rate.” <a href="https://twitter.com/om/status/1690903567551586304?s=20" rel="nofollow noreferrer">https://twitter.com/om/status/1690903567551586304?s=20</a>