Let me translate this blog post out of corpspeak...:<p>OpenAI has agreed to use Azure exclusively and for free, in return for Azure massively scaling up it's GPU capacity (which OpenAI was hitting before, hence threatening to move). In return, OpenAI gives an extensive license to Microsoft to use ChatGPT in Bing (and other products).<p>Now why couldn't the blog post have just said that rather than hiding all the details?
> <i>In pursuit of our mission to ensure advanced AI benefits all of humanity, OpenAI remains a capped-profit company and is governed by the OpenAI non-profit. This structure allows us to raise the capital we need to fulfill our mission without sacrificing our core beliefs about broadly sharing benefits and the need to prioritize safety.</i><p>All this time, I was entirely unaware of this.
Microsoft is handling the AI revolution extremely well.<p>They see the potential brand damage that comes with a such a product, so they probably just hold less than 50% of the company, to be able to say "this is not us".<p>Which then allows openAI to experiment and wow people, while google is anxious about showing 10% of its crazy AI capabilities to avoid potential PR disasters.
If I'm understanding this, the Codex model like code-davinci-002 and limited to 10-20 requests per minute on OpenAI, is now "generally" available through "Azure OpenAI Service" BUT<p>- you still have to apply to <a href="https://customervoice.microsoft.com/Pages/ResponsePage.aspx?id=v4j5cvGGr0GRqy180BHbR7en2Ais5pxKtso_Pz4b1_xUOFA5Qk1UWDRBMjg0WFhPMkIzTzhKQ1dWNyQlQCN0PWcu&culture=en-us&country=us" rel="nofollow">https://customervoice.microsoft.com/Pages/ResponsePage.aspx?...</a><p>- you will be rejected if you use a personal email address<p>- you will be rejected if you are not a "managed customer"<p>- you will be rejected if.. Microsoft thinks you are a real competitor to any of their products?<p>So my site aidev.codes is will be dead in the water if I can't navigate these things or afford the "managed" part. What is that? I don't know what a "managed customer" is but I assume its very expensive.<p>I mean, this is really the opposite of Open.<p>Or am I missing something? Is it really generally available and they took all of that stuff off?<p>BTW I am looking for an investor: <a href="https://news.ycombinator.com/item?id=34494480" rel="nofollow">https://news.ycombinator.com/item?id=34494480</a>
For people more well versed in this space: does GPT / OpenAI more broadly have a meaningful moat? It seems like there will be a number of these models out there and this isn't as great as say Google's up-till-now advantage in search relevancy.
> Azure’s unique architecture design has been crucial in delivering best-in-class performance and scale<p>Is there something unique about Azure, which makes it well suited for AI?
Related MS announcement with some more details: <a href="https://blogs.microsoft.com/blog/2023/01/23/microsoftandopenaiextendpartnership/" rel="nofollow">https://blogs.microsoft.com/blog/2023/01/23/microsoftandopen...</a>
> We’ve also worked together to build OpenAI’s technology into apps like GitHub Copilot and Microsoft Designer.<p>Good. I get better results from chatGpt than from Copilot. For example, often Copilot would invoke a bash command for copying a file in my python script, instead of using pythons buildin shutil.
I asked ChatGPT to write a poem based on this post. Here's what it wrote:<p>---<p>Two hearts, two minds, working side by side<p>In pursuit of a mission to make AI safe and wide<p>Microsoft’s investment allows us to stay true<p>To our capped-profit company, our mission, and you<p>From cloud computing to real-world use<p>We collaborate to ensure our systems are not abused<p>OpenAI API, Azure’s OpenAI Service, and Designer too<p>We’re working side by side to make our visions come true<p>Together we review, together we synthesize<p>Together we build, together we realize<p>The potential of advanced AI that’s powerful and safe<p>Ensuring its benefits can be shared far and wide
> So, we’ve partnered with Microsoft to deploy our technology through our API and the Azure OpenAI Service — enabling enterprise and developers to build on top of GPT, DALL·E, and Codex.<p>I wonder how much of choice will enterprises or large customers have in deciding what data to not send to Microsoft to train their models. If OpenAI is tightly integrated to all Microsoft products this may be a real concern.
This feels like an inflection point. Microsoft obviously sees huge upside, they're forking over a lot of value for this, and they must be expecting huge growth. Maybe Bing will be an amazing search engine in a couple years...
> Azure’s unique architecture design has been crucial in delivering best-in-class performance and scale for our AI training and inference workloads<p>translation: Microsoft is not Google (who wishes we'd explode).
The WSJ article on this announcement states:<p>>"OpenAI was in talks this month to sell existing shares in a tender offer that would value the company at around $29 billion, The Wall Street Journal reported, making it one of the most valuable U.S. startups on paper despite generating little revenue."<p>Can anyone say why they wouldn't have gone that route? Is that valuation possibly based on market condition from a year ago?<p>[1] <a href="https://archive.is/slxcE" rel="nofollow">https://archive.is/slxcE</a>
> Azure’s unique architecture design has been crucial in delivering best-in-class performance and scale<p>Is there something unique about Azure, which makes it well suited for AI? (vs AWS or GCP)
“Learn to code.”<p>That’s what folks were crassly saying to truck drivers, baristas, and cashiers.<p>The writing was/is on the wall: AI is only going to get more powerful and able to be applied to more and more complex tasks.<p>The thought was that “unskilled” labor would be the lowest hanging fruit, that automated AI - with some (but minimal) human oversight - would replace serious chunks of the workforce in various minimum-wage and “blue collar” sectors.<p>Machines don’t necessarily need to sleep, they don’t have labor unions, or laws that require healthcare or overtime pay. They don’t get upset, take things personally, seek revenge, or reciprocity like a person.<p>Sounds like that could be a threat to many kinds of jobs, many of them “bullshit jobs” (in the words of Graeber), but others as well.<p>It seems (to me, at least) that the more immanent threat is text-based AI - with some human oversight - replacing large swaths of the tech workforce (many of whom were leading the narrative about the truck drivers).<p>The incentives for companies like Microsoft, Amazon, Google, and others, to go this route are obvious, beyond the stated reasons why “low-skilled” labor is at risk. They already have enormous investments, acquisitions, projects, established platforms, and infrastructure related to AI.<p>I expect to see more partnerships like the one between Microsoft and OpenAI, from all of the major tech companies.<p>I also believe the connection between the acquisitions/partnerships and the mass layoffs will become more and more obvious as these acquisitions/partnerships continue to happen.<p>I can’t be the only person that is noticing this…
This technology is powerful and disruptive enough to merit social oversight. We get caught up in the wonder of the novel mechanisms and philosophical wanderings, whereas those who have the means to field this technology are effectively creating <i>facts on the ground</i> as we muse about it.<p>As just one dimension, let's briefly review education limited to testing. Options here range from (paid) cooperation between institutions and companies, to marketplaces for detection technology and testing systems, to complete rethink of testing at the institutional level, to regulatory intervention.<p>The dynamics and outcomes resulting from considering the above <i>before</i> or <i>after</i> widely available and improving LLMs are a fact of life, are possibly very different. We haven't yet crossed the Rubicon here and not sitting as captive spectators in a theater of ultra rich remains an option.
> This multi-year, multi-billion dollar investment from Microsoft follows their previous investments in 2019 and 2021, and will allow us to continue our independent research and develop AI that is increasingly safe, useful, and powerful.<p>I laughed at independent research. Why do they even bother..
Google recently said they are recalibrating the level of risk they're willing to take with releasing this technology. OpenAI's CEO responded saying<p>> "recalibrate" means "increase" obviously.<p>> disappointing to see this six-week development. openai will continually decrease the level of risk we are comfortable taking with new models as they get more powerful, not the other way around.<p>How does this resolve?<p>OpenAI not releasing something isn't going to stop Google or another player from releasing it. Won't they have to recalibrate too or be overtaken? Pandora's box is already open.
Somewhat off topic, does someone know how bing might integrate chat gpt into search. Is it to understand the prompt and filter results. Taking the question and summarizing it to search the index. Is it to summarize all the documents into an index and search that. Or to just be like chat gpt is now and use it to generate new results from it's knowledge base? I'm trying to connect the dots between a generative form like these are and how it would influence search in the future. Or is the lucene style index search on it's way out in a generative world?
'Extend'<p>Even capitalized and straight from the horses mouth, but as unsurprisingly predicted in: [0]<p>> * OpenAI will gain further investment...<p>We already know about ChatGPT Pro since that is a paid version of ChatGPT coming soon, meaning that the second prediction in [0] as also turned out to be true:<p>> ...ChatGPT by then will become a paid service<p>Now we wait until the startups depending on the service start to raise their prices.<p>[0] <a href="https://news.ycombinator.com/item?id=34201706" rel="nofollow">https://news.ycombinator.com/item?id=34201706</a>
Microsoft being late to the party is going to cost them a lot of money.<p>Google and Facebook are the clear leaders in AI and they been spending a lot of money building out their AI infrastructure over the last decade. You haven't heard much about Microsoft's investments other than them buying rights to OpenAI models. Seems like they could have saved a lot of money if they started investing years ago.<p>Google bought DeepMind for 400m 8 years ago. You could argue DeepMind is on the same level or above OpenAI.
The post states:<p>>"Azure’s unique architecture design has been crucial in delivering best-in-class performance and scale for our AI training and inference workloads."<p>Can someone say what is unique about the Azure cloud architecture here vs other cloud vendors?
GTP3 can generate some good press releases.<p>I presume that by integration of the Open AI services into the bing and MS Office, Microsoft’s mission (empowero every person and every organization on the planet to achieve more) makes lot of sense.