Today on the podcast, we're going to be talking about a recent acquisition that OpenAI has made, leaving us wonder, you know, why did they acquire this company? Was it something they were able to do before? Was it something they wanted to scale up really quickly? We're going to be talking about that.
In addition, we want to talk about the fact that it looks like chat GPT may be winning over sack overflow. And we're going to dive into that as well. So without further ado, let's jump into it. So OpenAI, of course, famous for chat GPT recently announced that it is acquiring global illumination.
This is a New York based startup that essentially uses AI to craft creative pools, digital experiences, and infrastructure. The acquisition is the first for OpenAI in approximately seven years. This is a really big deal, right? This is the first company that actually acquired and details regarding the transaction terms are still undisclosed.
We don't know exactly how much this was acquired for, but in a statement on their blog, open eye said, quote, the entire team has joined open AI to work on our core products, including chat GPT global illumination is the brainchild of founders Thomas Dimson, Taylor Gordon, and Joey Flynn. And it has it has engaged in a whole bunch of projects since 2021 when it was first created and notably the venture received support from prominent VC firms such as paradigm, benchmarks, and slope. And aggressively, the minds behind global illumination had earlier played pivotal roles in product design and development at tech giants like Instagram, Facebook, YouTube, Google, Pixar, and Riot games. So while serving as the director of engineering at Instagram, Dimson significantly contributed to enhance the platform's discovery algorithm and his contributions at Instagram include the initiative of teams focusing on the explore tab feed and stories ranking.
So he also worked on IGTV and a couple other things, right? So obviously he's very well ingrained in tech and did a lot of big moves in some big products on some big platforms. So then that begs the question, right? What is, you know, what is global illumination is actually going to do?
So one of the latest offerings from them is actually Bios, which is an open source sandbox MMORPG rem, it's reminiscent of Minecraft, but crafted from web platforms. So the game's future remains really ambiguous and given the resack position as possible that global eliminations endeavors under open AI, some umbrella might, you know, veer away from entertainment focused projects and more towards other things. So the open AI has previously steered clear of acquisitions. I think the AI giant right now, they have a lot of money in venture capital from Microsoft and VCs.
And so they have been proactively funding and offering grants to different AI firms over the years. And I think evidently AI is strategically positioning itself for a significant commercial breakthrough. So despite GPT for his worldwide usage right now, I think open AI reportedly funneled over $540 million into development last year. So it's definitely spending very, very heavily to improve its core products.
I think the investment also kind of encompassed the strategic hiring of talk talent from competitors like Google, which, you know, obviously is not going to be super cheap. And in addition, while open eyes revenue touched $30 million in the year before CEO, some ultimate has reportedly informed stakeholders about ambitious plans aiming to scale this figure to 200 million in the current year and envisioning a really robust $1 billion the following year, which could be impressive. But in order to pull this off, they're definitely going to need a lot more help. And it would appear that this new company they've acquired is going to help them do just that onto another piece of news, the fact that it would appear that chat GPT is beating stack overflow.
So if anyone remembers where this was at the beginning of the year, stack overflow is the website that developers use to post questions and get responses on how to do code projects, how to develop different types of code to code snippets. Now, the problem is stack overflow saw a massive decline in traffic after chat GPT launch, but people realized they could ask chat GPT for questions about coding to write code and all sorts of things. Stack Overflow took a massive hit in recent just recently, stack overflow has made a lot of changes to settings saying that they are going to charge people to to essentially use their data. They're going to train models because I believe Gigi BT like trained with using stack overflow data, which is, you know, pretty tricky for stack overflow.
And in addition, I believe they're training their own AI model for developers. But with all of that underway and all those announcements, it would appear in the, you know, the today's term that chat GPT may still be winning. So even with all of the super experienced data that stack overflow has, and I think it still is a really big tool data from analytics from may said that the platform's traffic declined on average 6% in every month of the year, culminating to 17% a 17% drop by April, right? So we're just seeing month over month, stack overflow is traffic is getting cut down due to chat GPT.
So a study at Purdue University titled who answers to be better in-depth analysis of chat GPT and stack overflow answers to software engineering questions. It's kind of a long title, but in any case, the study suggests that they found a concerning trend, a notable chunk of chat GPT's answers to coding related queries were either misleading or downright false. However, in a comparison test, 40% of participants favored chat GPT's responses and were swayed by its really good response style. So of 512 chat GPT responses analyzed, a whopping 52% were flagged as incorrect and alarmingly 77% of those preferred by participants were inaccurate.
So despite really glaring inaccuracies and chat GPT's answers, a subset of participants still lean towards them over stack overflow and stack overflow responses. So Semiticabir is a co-author of the study sets shed some light on the saying quote, participants overlooked the inaccuracies when they found chat GPT's responses insightful. Confidence, articulate, delivery style, even if misleading fosters user trust. So according to Kabir, the blend of this language really makes people like it.
So while language models like chat GPT promise really quick answers, they are not always 100% accurate. So stack overflow recognizes these pitfalls and has prohibited chat GPT derived responses. You can't ask coding questions on chat GPT and post some on stack overflow. So a lot of people are doing that right at the beginning of chat GPT.
And I think that this is really interesting, right? Humans actually prefer it. Now, the other caveat I will say is a lot of times when you hear these kind of articles and studies and they're like, oh, my gosh, 52% of the coding responses were inaccurate. That might be true, right?
If someone says, and also let's all define inaccurate, but you might say, like, hey, code me something that does X, Y, and Z. It spits out some code. Now a respondent might say, okay, this looks great. Pop it in to wherever they're developing and then they're going to go debug it.
They're going to fix whatever's wrong. If you're kind of a semi-experience developer, you're like, I just need something to throw something together for me real quick. I know it's going to be like referencing libraries that don't exist, whatever. I can go fix those and change those.
And stack overflow might be more cleaner code that you could have instantly copy and paste. But perhaps it's harder to find the result in stack overflow. I'm not 100% sure, but what I will say is on the project I'm currently working on, my developers literally say they cannot live without GPT-4 and that they use it every day. So, and also that they have like went away from stack overflow a lot.
So like I see this in real time playing out across real companies. And I do think that, you know, a lot of these studies are going to be saying, like, so it's funny because the takeaway from the study is like, Chai-Chipped is winning. People prefer it. And then everyone's like, yeah, but it's like less accurate than stack overflow.
It's like at the same time, sometimes all it needs is a few tweaks to what Chai-Chipped is coming up with and it can come up with it a lot faster and it can do a lot bigger components than stack overflow. So, I think that's a lot of benefits. And I don't think people should get, I think inevitably we're going to get less caught up about like, oh, it has like errors and be like, oh my gosh, it wrote me like an entire page of code instead of like two snippets that were perfect from stack overflow the entire page needed a couple of tweaks. So, well, you could say it was an inaccurate entire page in two seconds it wrote an entire page that I would have to piece together with a bunch of different things from stack overflow.
So, there are some pros and cons to it, but I definitely know a lot of developers that can't live without it. And so because of that, I think that Chai-GPT is only going to get better and I think stack overflow really needs to up its game and it needs to get its AI assistant out there ASAP and scale it because that is really where the space has gone. I think right off the bat, their biggest move is to be felt threatened by Chai-GPT was to replicate it with their own co-base. It feels like they just recently started doing that.
And so, I mean, for their sakes, I hope that they get it done sooner than later, but it's going to be an interesting space to follow and see if once they do create something, which they've announced they will, if they're able to actually compete against Chai-GPT or if the user, the brand name's so big and everyone is already using that by that point. So, definitely an area to follow in the future. Thank you for listening to today's podcast episode. Breaking down how AI is impacting your industry.
Today's episode is sponsored by AIBox, a no-code, AI app builder and marketplace, which just launched a crowdfunding campaign. If you are interested in investing in a new AI startup, you can go to republic.com slash ai-box. I'll leave a link in the show notes as well to learn more. The minimum investment is $150 and the maximum investment is $100,000.
Until next time, have a fantastic day.