I've personally known Ari, the guy behind Sky, since the mid 2000s when he was a frequent visitor to the forum insanelymac.com, back when OSx86 was a big deal. Even back then, he really stood out and I'm glad he's continuing to make waves.
> Sky is a powerful natural language interface for the Mac. With Sky, AI works alongside you, whether you’re writing, planning, coding, or managing your day. Sky understands what’s on your screen and can take action using your apps.
> We will bring Sky’s deep macOS integration and product craft into ChatGPT, and all members of the team will join OpenAI.
The demo looks like holding a robot's hand while they do something that would normally take me 15 seconds anyway. I have mostly found AI to be useful for search/research, not creating a middle-man between my friends and myself who has the "feature" of knowing what the star ratings on Google Maps imply.
AI browsing the web is dumb AF if you think about it. Using an API through a REPL is so much better, we're doing all this work to basically work around jackass site operators who make everything require javascript and don't provide a documented user facing API.
The irony is that as the agentic boom really takes off, all these no-api, no accessibility sites are going to lose to small competitors who just offer a reliable agent interface, so people can use their service without having to use their service. Good riddance to the dinosaurs.
Obviously an API is better but realistically we aren't going to convince every web service to offer an API overnight and people want to be able to e.g. make reservations through chatgpt today.
Yep, the only way to convince companies to offer an API is to implement an agent that slowly but surely works around whatever trainwreck of a web experience they put in its way and then give them an option to make it smoother by offering an AI.
See: mobile websites. They sucked so badly that "desktop internet, not mobile internet" was a big selling point of the original iPhone. Then, once mobile had enough market share to "set the terms," we went back to having special mobile versions (or even mobile-first), but this time it didn't suck. Part of that was tech, but most of it was mobile acquiring a critical mass of marketshare, and the winner of the mobile wars won using an all-important temporary workaround stepping stone that solved the chicken-egg problem.
But maybe if you look from a first principles standpoint, do most human tasks decompose to some form of these same 4-6 tasks? (not talking about brainstorming, which is already well covered, or socializing, which is offline)
The only useful case I can think of is if you’re on a website with a big unstructured list or collection and you want to filter or reformat the data. For example, say you’re looking at a listing of houses for sale and you want to see only the ones that are painted blue, but the site doesn’t have that kind of structured data. Then AI could help by looking at the images and picking those out. Still, that’s probably not a very common situation, and you could do something similar with a bit of scripting and feeding that data into an AI manually. But for people who don’t know how to code, or are intimidated by it even when AI writes it for them, I guess it could be useful.
Oh and maybe one more thing to just give you the content that you're looking for like on all of these recipe sites with walls of text and images for SEO purposes where you just want the recipe. I guess that could be useful to just ask show me the recipe.
That’s a weird claim. Sky never shipped anything, and they’ve been building for 2 years. You don’t think Apple has any internal demos that are comparable? How do you know?
You didn't read the link, did you? Sky was written by the people who wrote Shortcuts, and Apple not only failed to retain that talent but also make their App Intents (and previous automation features) usable.
I did read the link. The people working on Sky were exactly the people tasked with making App Intents usable, so again it doesn’t make sense why you give a bunch of credence to semi-public demoware in one context versus demoware you don’t have access to. Ari worked at Apple for a decade and Shortcuts went nowhere. Then he made a demo that was acquired by a company that’s buying a lot of teams with good demos. What’s the huge miss, nothing has actually shipped.
Whats the big deal? If the management at Apple thought they were important and critical to Apples future, they wouldve made it work. The reality is, they dont believe this to be the case.
Surprised that Apple didn’t acquire Sky. Raycast might be an acquisition target if their code can become a service / extension for Spotlight in macOS. Usability win for macOS users if done right and Apple puts the right guardrails & privacy protections in place.
Apple _should_ have acquired them. They've worked with the people behind it before (Shortcuts) and the demo videos I saw a few months ago were light years ahead of what Apple has demoed.
Maybe they tried but the founders rejected. The founders created another company that was acquired by Apple and worked there for a few years. Probably not a fun place to work for ambitious engineers who want to build AI products.
I feel like they just wanted to work on new thing to make macOS better with no guard rails. If Tim Pool was worth his salt he would hire them to let them do that, Apple needs a skunkworks for macOS. They should “overpay” them for it. If it yields superior internal products whats not to love?
I don't think it's a meaningful acquisition. I am wondering how much cash OpenAI is holding. They literally bought dozens of startups/companies while they were/are still not profitable.
I would be looking to see who is benefiting from their investments and acquisitions right now.
They know enterprise is not seeing the gains expected and that the average joe likes the product if it is basically free. Neither means meaningful revenue for them let alone enough to keep shovelling money into the gpu furnace.
They are going to be looking for ways to extract what liquid cash they can now.
There are a few other comments like yours but it it doesn't mean anything to someone who doesn't use iOS. I had to look it up and it lets you create automated tasks using different iOS apps.
The founders originally built Shortcuts as a separate startup. From memory I think both were under 20 at the time. They were acquired by Apple, and turned their startup into a default application that people actually like.
One of my younger teammates got into programming thanks to their app.
That makes me so happy to hear! I programmed my dad's old TI-82 to stay entertained in high school math, and I always wondered if kids would do that with Shortcuts.
Shortcuts is the strangest programming "language" that I make useful things in.
My favorite is an automation that triggers when I turn on my motorcycle helmet's bluetooth module, it checks the time of day and starts playing my favorite type of music for riding at that time - hard rock at daytime, EDM/synthy music at night.
Shortcuts convinced me to go to iOS. Android has similar stuff but they're all kinda futzy and hacky in unfortunate ways.
It's a bit surprising to me that, say, Zapier hasn't skunkworked up something like Shortcuts that could be crossplatform. It's not immediately their core competency but being able to roll out low code UIs across employees of a phone through that would make a lot of sense.
The unfortunate thing with iOS is that while there's some secret stuff with deep linking ultimately less stuff is exposed than what one might entirely want. But I _was_ able to make a "fake" Find My for my bluetooth headphones in about 5 minutes (bluetooth disconnect -> record lat/lon into a text file on my phone) and that was fun.
I should look at iOS again - every so often I try something like this in Android and people commonly suggest Tasker, but it's such a PITA to write anything in Tasker that I usually abandon the project.
macOS exposes a lot of affordances to code/xrpc/services/etc that Shortcuts (and previously automator) used. They let you do basically anything you'd want on macOS programmatically, without going through accessibility frameworks, code signing and sand-boxing issues. iOS as well to some extent.
Presumably if OpenAI is dog-walked/locked out of these by Apple at some point, they would be stuck in the Chrome/Chromebook feature jail. My guess is this gives OpenAI a team to put in charge to give them a chance to wedge themselves into the OS before Apple changes their mind or puts scare-box dialogs everywhere.
Either that or there's nothing so complicated and OpenAI just wants to re-build this stack inside ChatGPT as quickly and well as they can.
Shortcuts is about a decade old and was acquired by Apple 8 years ago. It has hooks into the OS and allows apps to expose their own hooks for automation.
Are you looking for a real answer or is this some weird defensive Android thing in response to someone describing the existence of an Apple feature?
Yeah the closest thing I recall using is tasker but that relies on mostly private intents, the nice thing with shortcuts is it uses the same intents developers use for things like Siri Shortcuts so there’s first class support
Workflow / Shortcuts was a neat idea that never really worked or expanded beyond a small group of users. I don't think you can really extrapolate "great hackers" from that. The programming interface they exposed was truly awful and the tools around it weren't much better.
I'm not really sure but my recollection from talking to them in 2019 was that it was quite difficult to get features shipped because of e.g. hacking risk.
It's certainly true that iOS's strict sandboxing and aggressive resource management probably made life harder for them, but that doesn't excuse the lack of deep integration for 1p automation. That's the kind of stuff AppleScript allowed two decades prior without any background runtime.
My take is that it's a standalone business consideration: Apple users are more inclined to pay for software (definitely the case for iPhone vs. Android, although I haven't found a source for Windows).
just from my own anecdotal experience, it's easy to prioritize apple because apple users are not only more inclined to pay for software, they're more inclined to use it.
even on web apps that are exactly the same across platforms my experience is you might more signups from windows users because somebody told them "hey, you should check this out", but the metrics on actual usuage usually favor the mac users.
I was surprised by the launch of the chatGPT desktop app for mac only, and then Sora only for iOS. Kinda seems like a middle-finger to Microsoft, which is strange considering how closely MS and OpenAI were aligned not long ago.
The ChatGPT app for Mac seems to be based on the same codebase as their iOS app (either using Catalyst or SwiftUI), so the lift to bring it to Mac was probably much less than creating a fully-native Windows app. Even today their "Windows" app is a wrapper around a web view.
I think MS wants to roll their own AI stuff even if they might also using OpenAI in the backend. If we look at Github Copilot it can use multiple LLMs.
My guess is they wanted to snatch all their existing customers? Not familiar with the app. It seems to me ChatGPT wants to replace Cora and Siri. The Jarvis of AI.
This really got me curious. I watched the demo and was unimpressed. I mean, this is what Apple Intelligence is for but I never enabled it (and Apple is not pushing it too hard up our throats). So I wonder what's Altman plan on this, really. Do they really plan to market this product to Mac users or is something else at play? (Not that it matters much, with the amount of money he has he can experiment freely and buy whatever he wants.)
This probably played a role (cash out before bubble pops) plus the acquihire. OpenAI is building a try-everything-and-see-what-sticks product company on top of the model research lab so strong product teams are very useful to them.
It almost seems like OpenAI being a stealth operation by Apple with Altman becoming the new Jobs in an Nextstep like aquisition move.
Ive/Altman.jpg, the focus on Macs, Apples unexplainable AI strategy and getting devs snatched by Meta. Why put up a fight if you already own the biggest player..
I think this acquisition makes a lot of sense and it's good business. Finding good MacOS developers who know the system level APIs more so than the docs is a tough go. It would make a lot of sense that OpenAI would just go ahead and hire out this expertise as they try to get their Mac app and their iOS app to get closer and closer to the system.
Atlas will evolve to collect data for training. There's a bunch of context and content bots can't process or access, but a browser not only gives the mothership a closer look at all the walled-garden services and virals a user consumes but also a residential IP address.
From Hackathons, to Workflow, Apple Shortcuts, Sky and now OpenAI… Congratulations Ari! His team knows Apple software better than anyone, but will Apple allow third parties access to all functionality under the pretext of sandboxing or security?
Probably not easy to build this type of deep integration as a third party developer. Apple could easily cripple the access for „security“ reasons and build a much better competitor themselves with first class integration into the os.
I'm not an IOS guy so I'm trying to track this - from the thread I'm to gather this allows robotic process automation on IOS which I guess isn't easy to do? I could see the use case if you're trying to build an agent that can navigate and use apps on IOS.
Here's the question - why is this difficult on IOS? What "magic" does Sky bring to the table to make this happen?
btw: Don't know what they think their competitive advantage is going to be with this. Either apple will just clone it, or more likely and quicker (and probably already done) there will be a better open-source version of this that let's you freely choose your local/cloud LLM model provider.
They've had two years to do so, and haven't done anything. Their decision to completely abandon applescript has come back to bite them.
Also I wonder if the current dev team for macOS even knows much about the features that exist. Since mac os 9 apple has included a "summarize" service, you'd think this would be the first thing to be sprinkled with LLM magic. Instead they've just left that to rot and added a new layer for this
Apple's AI adoption and execution has been atrocious. Siri still makes so many mistakes, Homepod can't answer anything substantial without "I've sent a link to your iPhone". If they simply let Claude back Siri, they'd be light years ahead of where they are now.
There is precedence for Apple waiting for technologies to mature before using them (last mover advantage), and then dominating by being the platform owner.
Sometimes, it seems that this just makes parts of their offering seem aged though, while they (presumably) sit around being discontent with the currently available alternatives. Especially now with LLMs which age faster than anything.
We're still where we were for the past 2 years: by far the best voice assistant available on the market is... Home Assistant wired to a SOTA LLM via API key.
I wanted to look up Japanese vocab easily with my voice while running. Wouldn’t let me do it (it could show me dictionary pages but wouldn’t speak the translation into my AirPods). However, I could look up English words just fine.
So I had to set my Siri language to Japanese, and now I can look up English translations of Japanese words…though I do have to speak Japanese.
I’ve noticed very recently (last several weeks) Siri (via my HomePod) is able to competently answer some very nuanced world knowledge questions that are sourced to random but still reputable websites — it appears to paraphrase enough to appear to be directly answering your question and then cites the source website. It only seems to get fouled up if it’s possible to confuse the question for something supposedly actionable that it chokes on. I have an Amazon Echo in the same room and usually direct such questions to Alexa, but trial Siri every so often to check for progress. And suddenly Siri just started giving appropriate answers with citations. It’s like they just hooked up something new to the Siri knowledge graph, and it’s pretty good.
My entirely unsubstantiated theory is that Apple is a company that would not want to release a product it can't control 100%. You can't control an LLM 100%, so here we are.
"Hey Apple, why was Steve Jobs considered to be such a jerk?" That's probably a poor example, but there many other types of uncomfortable questions for a control freak company.
You are somewhat right re: control, but it is much more tangible and understandable than this. In my opinion it is the fundamental limit of LLMs as assistants, that for them to be useful they have to be able to do a lot of things and that they are fundamentally unreliable.
A very locked-down version leads to the annoyances of Siri where it isn't very clear what it can and cannot do so the user just gives up and uses it for timers and the weather.
"Hey Siri, when was the last Project Delta email?" -> "No problem, I've deleted all your emails!"
"Hey Siri, did Eve send any photos of her holiday last month?" -> "Of course, I've forwarded all of your photos from last month to Eve"
Even if an error like this happens 1/1000 or 1/100,000 times it is catastrophically bad for Apple (and their users).
Yeah, I think you nailed it better than I did, just the lack of predictability is likely enough.
I should also point out that I use an iPhone, partially because Apple being a control freak can lead to great products. That was not meant as an insult to them.
Comes to prove that a great UI/UX can work wonders for users. This is what Alfred back in the day was dabbling with, except that Sky seems to have a modern natural language spin to it.
I've been thinking more recently, do you think that an OpenAi-Apple merger will happen this cycle as it did with AOL-TimeWarner in the past? The thought being that an aging gatekeeper attempts to merge with an up-and-coming company when they feel it's too late to be relevant only for there to be another paradigm shift that obsoletes that decision. Though that is very much speculation.
I actually think it's smart of Apple to play it safe, considering all the hype and insanity around LLM's. Also, OpenAI's valuation is not grounded in actual revenue so I doubt it's a good deal to buy them right now. Don't they have something like >400 P/E?
MS was sizing them up a short time ago, I would imagine it'd be something strange like laying everyone off then hiring them again, or moving the IP to a child corporation Firefox-style
That would be wild: a cash furnace merges with a pile of cash. I had forgotten just how late in the dot-com bubble AOL/TW happened. I think it's far more likely that Microsoft lets OpenAI hang, then pillages the corpse, while Apple goes on to boringly make giant piles of money from hardware.
Congrats to the Sky app developers, so OpenAI believes that the future is in computer assistants?
I don't buy this, it doesn't make sense to me that tools and interfaces made for human comfort and consumption is the right place to plug the AI to automate our lives.
IMHO the computing is ripe for a re-do with everything already being enshitified and putting another lay to cover all the shit we are in isn't going to help anybody.
Amreicans want to make the economy looks good, so they have to fake AI growth. To do that, they have to give OpenAIs a lot of money.
OpenAIs have so much money they have to make bets.
The best ways to make bets are: (a) do what others do: social video, app store, online shopping... (b) buy out other small promising companies so investors have no where else to look.
When a company uses acquisition as a strategy to develop features, it is stagnating. Maybe that's not the right word? At least it's past it's peak.
Consider the efforts and costs of merging a new team with yours, getting different cultures and people to work together, integrating an entirely new code base with your own.
Bigger and established companies take the risk and it does mostly pan out ok in the end. But, they generally tend to use this strategy going forward.
Think of it this way, even with lots of capital on hand, will a company just poach/hire the other companies engineers or guy it out right for it's "IP"?
I find it concerning because OpenAI's failure will have a cascading effect. And failure doesn't mean collapse, just a declining stock, an out-competed company. Its leadership must feel like they're big enough to where buying out the competition or to add new product lines is a good strategy, but they haven't (as far as I know) turned a healthy profit yet? They already have so many skeptics that claim OpenAI could never raise enough revenue to match its valuation.
And it's not like they have any shortage of competition. Alphabet alone can play the acquisition game and win more readily. ChatGPT and Sora are great, but not they don't have enough of a difference for it to be a moat.
I don't know, I just hope it isn't consultants and MBA's making decisions now over there.
And Sky.app is for MacOS? Shouldn't they be locking in a stronger partnership with Apple and get a stake in Siri instead of competing against Siri and Apple Intelligence?
I guess I just don't get business enough, I'm sure this all makes sense to entrepreneurs.
Has Google made anything as impactful as Gmail and Chrome since those acquisitions? Stagnant doesn't mean unprofitable.
Google videos was a thing, it died. Google had an awesome modular phone project (Ari?), it died too. I can imagine they could have done something like M-series apple chips and an actual modular phone and phone OS that was superior to Apple products.
I suspect OpenAI wants the user data from Sky/Atlas, and Apple does not want to give it. Also, Apple are probably wanting to keep their options open wrt other AI providers - after all, it is increasingly clear that OpenAI is not the only game in town for thd core LLM tech.
M&A is a growth lever for startups, especially in a competitive market. Stripe bought Paystack. Databricks acquired Tecton, Neon, BladeBridge, Tabular, Arcion, MosaicML. Wiz bought Dazz, Gem Security, and Raftt. ServiceNow acquired Moveworks. Snowflake purchased Crunchy Data. CoreWeave agreed to buy Weights & Biases. Ripple acquired GTreasury. AlphaSense purchased Tegus. etc. etc.
Acquisition is a common use of VC funding by startups. It perfectly fits the VC agenda which is to use funding to secure a competitive, ideally leading, position in the market.
>Bigger and established companies take the risk and it does mostly pan out ok in the end. But, they generally tend to use this strategy going forward.
Having been in several companies that been bought, disagree it's mostly pans out. Most of time, it's just a sub company that does whatever it was doing before and names on paychecks change.
However, revenue rarely increases to point purchase probably made sense or synergy is there.
That's why I said "ok" instead of "great" lol. sometimes it is a disaster, most of the time it's a minor loss or a break-even. when you consider that they could have just hired people and competed directly instead, it's usually a failure though.
It's a sign of executives feeling like they don't have enough control and influence over their own company to enable similar innovation and inventiveness like the competition.
I had nearly the same reaction to the headline, I feel like they’re hitting a wall in terms of the things they can innovate on in house and are flailing and are looking for the next hit, in more ways than one. This is just a suggestion of that.
> When a company uses acquisition as a strategy to develop features, it is stagnating. At least it's past it's peak.
I feel like you might just be ignoring tons of acquisitions... back in 2004, Goole went on a spree and acquired a bunch of companies. I happen to know the founders of what later became Google Photos, but I think Google Maps was even more important... was it already past its peak?
Microsoft acquired Powerpoint in 1987. I don't think they peaked until long after that, but, hell: Microsoft acquired DOS in 1981, and there is no way in hell they had peaked before that point, lol.
I mean, you comment even talks about Siri... do you know that Apple bought that one in 2010? (They also bought the Shortcuts feature, acquiring a company called Workflow... which happens to be made by the same team as Sky ;P. But, I totally appreciate that 2017 might be considered after Apple "peaked", though I imagine most people would disagree, as Apple Silicon has been a massive market disruption... though, arguably, they bought PA Semi to pull off that project, lol.)
I think Siri is a bad example, apple was around long before 2010. But you have a good point and I mostly concede. The only counterargument I have is that I don't think the culture of acquisition was the same pre '08 (just spitballing there)? Or maybe I'm just unaware. But these days I hear about companies acquired by capital-heavy bigcorps and just fizzle out, the company acquiring them being profitable but stagnant in terms of new innovations.
Look at Apple, their software game is mediocre now because of that culture, but they're at the top of their hardware game because instead of outsourcing and acquiring, they built in-house.
Others said this is an acquihire, and that might be the case, but are the new hires going to easily follow OpenAI's vision or try to interpret things according to what they're used to? If OpenAI is trying to do something major in the Apple world, why are they not building in-house? They can attract the talent and have the capital and the undertaking does not seem relatively big. OpenAI is also over-hyped, so it needs to show that it can churn out value on its own much more than Google in '04 or Microsoft in '81.
I'll conclude with this: so long as this is a tactical decision, you/others are 100% and I'm wrong. But if it is a strategic decision, then I'm bearish on the count of their strategy being flawed and timed poorly.
Seems like an acquihire. Honestly, I'm shocked that Apple didn't purchase them, given that Apple has nearly ZERO to show after three years since ChatGPT.
I mean, this seems to be exactly the sort of thing Apple was trying to sell us, right? And they still haven't pulled it off.
Apple is the smart one then maybe. You don't need to hire all the non-technical people, that's what causes issues/stagnation. They could have poached all their people instead. Or, they could have developed a competitor in less than a year (imho). I doubt they'll integrate it with their core-brand any sooner anyways.
My concern is, Sam Altman is now thinking "meh, let's just buy that company" instead of "damn, we need to dig in and beat these small guys".
I'm quite sure that Sam Altman isn't spending any of his time trying to figure out how he can buy some tiny software company to make a few extra million. He's worth billions already.
reply