AI Applied: Covering AI News, Interviews and Tools - ChatGPT, Midjourney, Gemini, OpenAI, Anthropic

OpenAI's Comeback with 5.6 and Work Tool

13 min
Jul 14, 20264 days ago
Listen to Episode
Summary

OpenAI announces a major comeback with GPT-5.6, a new Work tool for computer control, and improved voice features, directly competing with Anthropic's Claude offerings. The hosts discuss how OpenAI has caught up to Anthropic after falling behind, and explore the competitive dynamics driving innovation in AI pricing and product features.

Insights
  • OpenAI has successfully caught up to Anthropic's feature set after a period of perceived decline, particularly with Work tool matching Claude's computer control capabilities
  • Competition between OpenAI and Anthropic is intensifying on multiple fronts: model quality, pricing, rate limits, and feature parity, benefiting users through better products and lower costs
  • Users are increasingly adopting multi-model strategies, using both Claude and ChatGPT for different tasks based on speed, cost, and accuracy requirements rather than relying on a single provider
  • Product-market fit remains uncertain; companies are consolidating features into existing apps rather than launching separate browsers or tools, suggesting users want integration over fragmentation
  • Voice quality improvements represent a potential breakthrough moment that could drive mainstream adoption by making AI interaction feel more natural and accessible
Trends
AI model providers consolidating features into core applications rather than launching standalone products (Atlas browser shutdown)Competitive pricing and rate limit resets being used as direct counter-moves between OpenAI and Anthropic within minutes of announcementsMulti-model evaluation becoming standard practice among power users to identify strengths and weaknesses of different AI systemsSpeed and efficiency optimization becoming a key competitive differentiator alongside model quality and reasoning capabilitiesVoice interface improvements emerging as a critical factor for mainstream AI adoption and user engagementEnterprise focus shifting from job displacement narratives to practical productivity tools and workflow integrationBrowser-based AI features becoming embedded in existing browsers (Edge, Chrome extensions) rather than standalone productsRate limiting and token management becoming important considerations in user decision-making between competing platformsCoding and development tools becoming primary use cases driving feature development and model comparisonIntegration of multiple AI models into single workflows as a best practice for achieving optimal results
Companies
OpenAI
Announced GPT-5.6, Work tool, new voice features, and shut down Atlas browser to consolidate into Chrome extension
Anthropic
Competing directly with OpenAI through Claude, Claude Cowork, and strategic rate limit resets to counter announcements
Google
Mentioned for Gemini model performance and integration with Chrome browser for ChatGPT extension
Microsoft
Edge browser with embedded AI features discussed as competitive alternative to standalone browsers
Perplexity
Mentioned as competitor with its own browser offering in the AI search and research space
11 Labs
Referenced for voice synthesis quality as benchmark for evaluating OpenAI's new voice capabilities
AIBox.ai
Jaden's company offering model comparison tool for less than $10/month to test multiple AI systems
People
Jaden
Co-host discussing OpenAI's comeback, competitive dynamics, and sharing personal usage patterns across AI models
Sam Altman
Referenced for strategic decisions on product-market fit and messaging around AI's impact on employment
Dario
Referenced for public statements about AI's impact on white-collar jobs and market positioning
Quotes
"OpenAI is having a day, man, right? I mean, like it feels like they are coming roaring back, like kind of flying out of the gate with a ton of stuff."
HostOpening
"I love this because I think OpenAI or Anthropic has really run away for the last two months with Claude Cowork and everything that is capable of doing where it controls your computer screen."
JadenEarly discussion
"I think OpenAI might have won me back with a lot of these features here."
JadenMid-episode
"I don't like it when Anthropic is like by and far the leader and they're kind of crushing it. I mean, I'll go all in, but like, I don't like it because I want that kind of competition."
JadenCompetition discussion
"The entire conversation around AI is how we work. So there were the big conversations from Dario and from Sam and everybody else saying, this is going to take all the white collar jobs."
HostMid-episode
Full Transcript
Jaden, OpenAI is having a day, man, right? I mean, like it feels like they are coming roaring back, like kind of flying out of the gate with a ton of stuff. First of all, I'm hearing a ton about how 5.6 is working for people. That's pretty cool. It's being compared to Fable 5 essentially. I haven't done benchmarks myself, so I can't say. But then a bunch of other things that I wanted to touch on, Jaden, because I heard you talking about this stuff on some of your other podcasts too. First of all, is this called Work, I guess, which seems like a kind of a co-work sort of competitor. And then there's the voice thing, and this has been making the rounds. I keep doing it. In fact, Jayden, just before we got on today, I was like, hey, are you still the old voice or the new voice? Because people are raving about the new voice, but they said it's rolling out today. I don't have it yet, but I also got off the crazy max plan on OpenAI, to be honest. Anyway, your call, what do you want to sort of like start with on this? oh man i mean there's so much i think the thing that i'm really excited about is the new open ai work which is the competitor to claude co-work and i also like how similar we're we're naming things now yeah work co-work like we're not going with that though i'm kind of yeah yeah it's it's good at least it's like a comparable product and in your brain you understand what it is um but no i love this because i think open ai or anthropic has really run away for the last two months with Claude Cowork and everything that is capable of doing where it controls your computer screen. OpenAI was getting into, I mean, they've been doing this for so long. They've been doing it for longer than anyone with, if you remember, ChatGPT agents like a year ago, maybe a year and a half ago. And I remember, you know, in the early days testing it out, it wasn't quite good enough. It had a virtual private computer that it was running. It wasn't running on your actual computer. Then they came out with Atlas, which by the way, that's a whole nother piece of news, which is that OpenAI's Atlas browser that they launched is shutting down. So they're killing it and they're putting all the features from that into a google chrome extension chat gpt google chrome extension and into just a beefed up chat gpt app so i mean they're basically copying what anthropica has done they have a clod app so they're just having a chat gpt app they're like like you don't need a full browser i guess although yeah interesting but uh there's other people with browsers out there we still have perplexity with their browser out there so there's other browsers out there it looks like right doesn't doesn't microsoft doesn't copilot yeah edge has a edge has a bunch of AI features embedded into their browser that actually is pretty impressive they might have been one of the first I've been a big edge user in my life but I did download edge and start using it when they had uh loaded up a bunch of these really cool AI features so yeah I think it's it's a really interesting moment right now opening I feels like they finally caught up to Anthropic which I feel like they probably hate me saying that because you know they don't want to feel like they were behind but they've caught up um 5.6 is just as good as fable uh many people are saying online they now have uh work where actually can control your computer they're uh consolidated a bunch of software so i think they're in a really strong position right now and if i'm being honest like you i canceled my 200 a month um open ai account i have two claude max accounts that are 200 i probably cancel one of those we go one open ai and one anthropic for when the rate limits hit me i go back and forth between the few the two and they do have of different cool features So yeah I think OpenAI might have won me back with a lot of these features here I fascinated by it And by the way, it's a good moment to sort of like shout out the fact that we do keep going back and forth. And if you haven't tried AIbox.ai, this is the best way to do this without spending a gazillion dollars. This is for less than 10 bucks a month. Jaden's company here that has been going gangbusters where you can just like compare answers back and forth, which is what I do all the time. But, you know, Jaden, to be honest, I hadn't been doing it much of late. I'd sort of like lived in Claude. Gemini, I don't know. I know they say Gemini doesn't have a frontier. Sorry, they obviously do have a frontier model, but they haven't had anything in a little while. It's sort of falling behind. And it's kind of probably a different type of episode. But I still love Gemini. I use Gemini just sort of as default. But I've put my team on looking into... chat, which I think is really fascinating because you and I were talking for a while before we got on today just about Claude Code and Claude Cowork and how those are starting to come together, excuse me, starting to come together a little bit just in terms of their functionality. Cowork is getting, you know, I think stronger and more solid and skills is starting to really represent agents in a way, which is really interesting. And then Claude Code, I don't think You can take over your screen yet or anything like that. But it's just interesting. Cloud code is also getting much more user friendly. It feels like day by day without even announcements. So now I'm sort of tempted to do the same thing as you, which is to kind of go back over and open AI again, open ChatGPT again with 5.6, all that kind of stuff. It'll sort of like feel like going back and visiting your elementary school. It feels like that's how long it's been. But just sort of seeing what's happening on that. But Jaden, when you think about that, when you think about how you use Cloud Code and Cloud Cowork, you're building like a ton of cool apps. You're doing a ton of cool stuff with AI Box, everything else. What's the, for you, the value proposition of going back and using ChatGPT is sort of just to check or the other thing I keep hearing is, and then there's a back and forth race on this that you were posting about, but it feels like now they're starting to compete on price a little bit too, right? Yeah, they're actually competing on a bunch of different levels and I love this. I don't like it when Anthropic is like by and far the leader and they're kind of crushing it. I mean, I'll go all in, but like, I don't like it because I want that kind of competition. And like you mentioned, I made a post on LinkedIn where essentially opening, I made their big announcement. They're like, Hey, look, we just dropped the new version 5.6. And we also have the new work tool that can take control of your computer. And then literally 10 minutes later, Anthropic drops a post and they're like, Hey, everyone, we've reset all of the rate limits on fable five. We've extended the fable five usage. So you can just keep using that. And it's like all your credits are reset. Anyways, it was hilarious. because to be honest, right before that tweet went out, about 10 minutes before, I was at 92% of all my Claude usage. And as soon as that thing got to the end, I was just getting ready. I'm like, okay, I'm finally gonna make the switch. I hit my rate limit. I'm gonna go get Codex and we're gonna like get all set up in there. And then it reset and I was like, well, maybe in a couple of days once I use these tokens. This is really smart So I do love that I love the competition here And I think that we finally at a place where like the competition is going to get prices down but it going to get the products better Opening AI feels like they really put a big effort on refocusing on what a lot of people are using what important I'm getting actual work done. And so I'm really excited because of because of that kind of refocusing. It feels like we're seeing from opening. Do you feel like it's, you know, all of this stuff is like how we work, right? Like the entire conversation around AI is how we work. So there were the big conversations from Dario and from Sam and everybody else saying, this is going to take all the white collar jobs, thinking that that would probably boost their IPO prices, my guess. And then it almost, I think, did the opposite to a certain extent. Not that we know that price yet, but more that the backlash has been tremendous, right? As we sort of like know from back in the graduation speakers, everything else, when you talk to people, people tend to be pretty negative about AI. And now it feels like all of these things are trying to become better user interfaces, better, easier to use. I still feel like we're sort of on the frontier, Jaden, people like you and me and probably a lot of our listeners too. Like when you go out into the world, that's not how people are acting. But again, just having to use, getting to use code and not having to open the sort of like the DOS style terminal windows is huge. But I love what you said about the browser, like Atlas shutting down and things like that. Because I think what Sam Altman had realized early on with this too is that everybody's still trying to find product market fit. Like what do people actually want? I don't think they want a separate browser. I don't think they want Sora, a new Twitter. I don't think they want, you know, they don't want that. But what they do want is all these things integrated into one thing and make it super user friendly, which then brings me to the new voice as well. And this is, again, we keep testing it. It still hasn't come online and everything like that. But like when I see the demos, and look, I'm always a little skeptical of demos for sure. Sure. But even knowing where it is now and knowing how good 11 labs tends to be, if it really is like that, I see no reason, Jayden. And like where we are right now in July 2026, if this is where we are now, this is only, you know, it's exponential in how much better voice is getting. And that's where I feel like maybe a breakthrough starts to come in terms of one way or the other. It's either going to turn everybody and say I have forever or it's going to make people feel like, oh, now this is something I can really interact with, you know. Yeah, 100%. And there's a bunch of different pros and cons between the two. One interesting thing, I was talking to my wife the other day, and she was like, because like, I've switched everything to Claude, but we still have like a ChatGPT account. And sometimes I max out our Claude account, it hits the rate limits. Anyways, she was telling me the other day, she was like, I was like, trying to get this thing done for like the kids homeschool, help them make this curriculum that for this unit we're working on. And she was like, and like, it was driving me crazy, because I kept asking Claude to do it. And like, it would do it, it just took so long. She's like, then I would ask ChatGPT. And it was like, instantaneous she's like i love chat gpt it's amazing and they've added so many new features that claude only used to have like they can make all these do incredible deep research all the stuff that i wish they could do forever felt like claude nailed it they didn't it feels like opening eyes all back on board but anyways she was super thrilled about how fast that they were doing it so opening eye has a really um impressive algorithm on speed and efficiency and when you can and can to do that And Claude it feels like it just going to always maximize the you know it going to give it the maximum amount of like power but it also the most expensive And it going to take the longest Now if I coding a new feature in a website and it like a really complex piece of software, and it has to get it perfectly right, I'm gonna spend more money and say, take as much time as you want. Like, I mean, relative, right? Like what, who cares if it takes you two hours versus if it takes you 10 minutes, because I just need this feature to be done and it's going to help me for you know two years so at the end of the day like there are different pros and cons to the two different platforms um and but i've actually also seen with the new version of what open has come up for coding i've actually seen some people i mean this is just tweets on twitter so you know take this with a grain of salt but they're like i'm using codex right now and um chat gpt 5.6 to go clean up mistakes that fable 5 left in my code base so So like there are differences in these tools and sometimes putting both of them, like I always will take like Claude, like something I've been doing for the last couple of months is like, I'll take Claude and say like, hey, look at this website or the SaaS I have and do like an in-depth SEO audit, security audit, and like speed optimization on it. Like these are three things I do frequently. And I redo them over and over again as I add new features and things. Well, turns out if you get those same audits done by something like Codex afterwards, you actually get like some different results and a couple different things that it was that it missed that it was able to get and so there is like a really good value in using multiple models to do the same task over and over again and to give you two different perspectives on it yeah i i think that's why constant comparing of of models is great and sometimes it's as just as simple as i just like how claude just talks back i again i think gemini is still i know the benchmarks aren't there, but I still find it just for day-to-day stuff, just one of the smartest things out there. But then for deep work, I got to say, I'm kind of like not as into Gemini as I was before. But this is the point about keeping up. And the only other thing on price before we wrap on this is I was looking at, you know, sort of like how much I was spending. And I don't spend a lot on API credits or anything like that. But like, you know, but I have like an open, you know, a couple of open claw agents. And I don't mind spending money because I know that I'm getting value out of it. And I think just sometimes when people say like, oh, 20 bucks a month, Jaden, it's still like 90% or 89% or something like that on free ChatGPT accounts. It's just they're not seeing it yet. I'm just sort of like wondering what the business model for that is going to be. Anyway, guys, so glad to be talking about OpenAI again. OpenAI, welcome back to the conversation. I know you guys are working unbelievably hard. We love our friends at OpenAI, don't get me wrong. But I know that even inside there, they were like, oh my gosh, Like, come on, like, where's, you know, the oxygen in the room has all been sucked out here. So super excited to have them back in the conversation. They have a phenomenal product. And listen, y'all, we are getting closer and closer to our 200 review mark. Jaden and I have a giant countdown every day. We celebrate it. Where are we, Jaden, right now? We are at 185. Oh, guys. So close. You can do this. We can do it together. We're so grateful for you listeners, everything else. and always look forward to these pods with you. But just remember to leave a rating review and tell your mom too as well. Yeah, exactly. Guys, we will see you all on the next episode.