OpenAI and Jony Ive in talks to raise $1B from SoftBank for AI device venture
ft.comMan, I can't even begin to imagine what they're going to come up with. Titanium AIPods. Non-detachable alluminum RayBans. Google Glass, but it's a literal Black Mirror worn on your face. Self-driving Segway. Transparent plastic smart lunchbox. Electric long-board that's also a tablet. Fist-mountable selfdefense impact jackhammer. 5G base station Zurb ball. A keyboard that starts repeating nnnn when you think of what you want to type. Neuralink, but for discreetly listening to major record label IP. Contact lens payment system where you wink to tip your landlord. Quantified shelf. A car without a steering wheel, made entirely out of Corning glass. Heelies that auto-reverse when they sense an oncoming homeless person. Van-life, but it's a sentient executive jet plane. Longevity device that puts you to sleep for at least 8 hours every day. Intimacy fit-ring. Schedulable dopamine-inhibiting nanobots? This could be huge.
If you're this good when trolling for 5 minutes, I would be amazed to see you work seriously. Honestly some of those ideas are... terrible, of course, but awesome in their own distorted way. And also very Ive-like.
Bravo, sir.
This is hardly trolling. I just birthed 12 startups.
I have 12 checks ready to go.
This completely sells it
Just imagine a Ive-designed retina-scanning-orb! Easily worth a billion to SoftBank, easily!
Consumers will love it. Give them 50 NewBucks™ for their service and you'll get retinas by the billions.
I need a self driving Segway, yesterday.
Should be able to avoid cliffs without a radar. You can always add it back later anyway.
An electric car that you need to turn up side down to charge? An AI assistant device that looks like a trashcan?
I came across this video of a Tesla that drives upside down: https://youtu.be/Qq5Q1qKW-1g?si=RVDHJZm47s-rCot-
Bit of a guilty pleasure, but it’s a sight to behold for sure.
> Heelies that auto-reverse when they sense an oncoming homeless person.
JFC lmao
Self driving trashcan, brings your trash to the landfill while you sleep so you don't have to pay for garbage pickup. Enable the automatic trash sorting attachment for only $999/month.
You should visit https://www.halfbakery.com/
Many fancy smart fridges are already quantified shelves. I see you are still adding stuff... let's keep it going.... Rideable Boston Dynamics Spot for commuting. An smart vape-pen line of medications. Smart Wigs. Self driving strollers.
Woah, quantified shelf market is done for. Point taken. Idk, Boston Dynamics Zoo? How is that more ethical though :$
Titanium Neuralinked Yarmulke
> Heelies that auto-reverse when they sense an oncoming homeless person.
If you brand these ‘Wheelie Safe’ and market to upper middle class urbanites you’ll have more money than you know what to do with. Careful, you might end up a billionaire!
Plot twist. He is Ive.... or a very good ai representation.
Yes, you CAN begin to imagine what they're going to come up with!
Haha truly a creative mind to come up with this
Non-detachable aluminium RayBans. You’re welcome.
Let's stay constructive and just go with sheet metal for V1 then ;-/
What the hell is an "AI device"?
Unless we're talking robotics, I don't see why you would need a separate device for AI. Seems like we already have all the right form factors (phones, speakers, etc).
This reminds me of when Facebook tried to make a Facebook phone. Turned out iPhones and Androids were significantly better as phones and they could do Facebook just fine.
The very real leaks that I’ve very really seen suggest it’s a bit like HAL-3000 with an upgraded bezel-less Retina display red orb that goes all the way to the diamond-cut chamfered edges, but when you ask it to do something instead of saying “I’m afraid I can’t do that, Dave” it starts playing latter-day U2 albums.
> it’s a bit like HAL-3000
Dave, there is only one HAL3000.
> What the hell is an "AI device"?
A phrase used to extract a billion dollars from idiot investors.
I always love reading myopic technology takes on this site.
Based on the article, it sounds like they're in exploratory talks to figure out what that "AI device" could be. I agree that it's likely we've already figured out the right form factors for such a device, but it's not a leap to imagine a significantly better UX than what we have from today's computing devices.
The movie "Her" seems like a decent blueprint for a post-smartphone "AI device".
There was this device in Neeuromancer. Like a small stone, you touch it and can talk to AI in your head. That’d be awesome.
Ya. Even a limited use AirPod-as-language-translator-in-real-time could be a good use of technology (something like this was in Hitchhiker's Guide and many other sci-fi). The leap in AI skill to trust translations more could enable the leap from looking at a phone to just having a bug in your ear.
Interestingly there's already an AI powered earbud [0] on the market that claims to do that; whether it's any good or if OpenAI can do better remains to be seen.
But how long before it's capturing audio from your conversations to use as training data?
The babel fish.
If you reverse this you could have your offshore call center sound onshore in a matter of one OpenAI sign up.
> What the hell is an "AI device"?
Puck and a pair of AirPods? (Maybe compute in the case?)
A device optimised to run a set of LLMs and not sacrifice battery for e.g. a screen seems like it would have a niche, if it just worked. Picture the AIs profiled in the later series of Westworld.
Would it not make more sense to use OpenAI LLMs to be hosted away from these device. It would also align with Jony Ive's focus on design over everything else. Yet another data consumption device?
> Would it not make more sense to use OpenAI LLMs to be hosted away from these device
The lag is annoying. I don’t remember Ive’s position on privacy, but shipping everything off to the cloud isn’t great for that.
There might be satisfaction in having a childlike AI running on device who can call grandma—or whatever-when it needs help. One could tune it to different levels of spontaneity (“that’s a Magpie, in case you’re interested”) and proactiveness (“I noticed a new restaurant nearby and you haven’t eaten all day, so I grabbed a placeholder reservation”) that shouldn’t require a lot of phoning home to the mothership.
If they can come up with an Alexa/Siri kind of experience that doesn’t make me want to throw it out the window, I can see the separate device not being too hard of a sell, while at the same time avoiding being at the mercy of whichever device provider they would have to play along with otherwise.
Of course Alexa is getting a LLM as announced recently in the next few weeks. I don’t think Amazon and Apple are going to miss the bandwagon.
On the Apple front they added a very small LLM to the keyboard autocorrect that is better at semantically correcting and predicting what you’re writing in iOS 17. It’s a ducking lot better already, but I imagine over the years it’ll round out to be incredibly useful.
Do know the margins to be made on a Prime Radiant? Bullish...
Maybe it will be something like Humane’s AI Pin
Something something razor thin aluminium.
On a serious note, I wonder if Altman will be able to control Ive’s obsession with form over function like Jobs did.
elegantly chamfered* aluminum
Don't forget rounded corners.
$1Bn pre launch funding + Softbank = not a good omen
I don't know much about softbank, but my brain strongly associates them with bad bets, failure, and losing money.
Add to that weak cell coverage, exorbitant fees, and a privacy nightmare of jumbled up offerings (Yahoo! Japan, PayPay, Line, Telecom, ISP services, etc.), and you get the full picture.
SoftBank is originally one of the big 3 telecom providers in Japan. We do joke around here about our phone bills increasing anytime Masayoshi Son makes a bad investment.
Too much money skews everything in product development
There's a reason why the iphone team was so small (https://www.theregister.com/2014/03/27/apple_employee_reveal...) you need to be ruthlessly focussed and not pressurised by outside investors or managers who don't know what they're talking about (Softbank)
I'm not sure what the money would be spent on at OpenAI. Maybe they will keep it small, and hyper-focussed. But I worked in VC for a while and VC money was described to me as rocket fuel, you don't want to go dousing it on things before the rocket is built because it stinks and has a tendency to blow things up
They invested in Alibaba about 20 years ago, made something like 2000x their investment ($20M to $50B), and have been wrong about everything since.
They were an early investor in WeWork, and paid Adam Neumann to go away instead of throwing him out head first.
For hardware devices that raised >$100mm before launch, how have they gone?
Successes:
* Tesla
Non-successes:
* Magic Leap
* Juicero
* Nuro
* Segway
What else am I missing? I suspect there are other successes that raised >$100mm pre-launch that I'm not thinking of.
I wonder why they wouldn't start smaller. That said, I still expect they'll succeed, and I hope they succeed. I don't believe in any rules of thumb, great startups always defy the odds.
According to public sources Tesla has raised 'only' 60.5m before delivering the first roadsters in February 2008. Though that total increased to 100.5m by the summer of that year. But I get your point.
[0]: https://www.startupranking.com/startup/tesla/funding-rounds
The original roadster was a Lotus Elise with laptop batteries. It was essentially a prototype, albeit a production prototype.
and it kicked ass
Indeed. But partially because the Lotus Elise kicked ass :P
A bunch of VTOL/eVTOL and space startups also probably raised >100$ pre-launch I guess?
Skydio raised way more than 100 million and I'd call them success, although I'm not sure if they raised that much pre-launch.
We raised ~$60m before R1 launched in 2018. And we were still under 100 total when S2 launched at the end of 2019.
I'm not sure how things would have played out if we had 100m+ before we launched anything. We may have delayed launch even further and missed out on valuable market feedback and brand awareness.
Oh man Juicero. Still remember watching someone tear one of those down. Absolutely incredible amounts of wasted work, materials, ludicrous overengineering to squeeze a damn bag.
That was AvE of course: https://www.youtube.com/watch?v=_Cp-BGQfpHQ
If you had invested $1B in each of those 5 companies, you would probably still have made a great return, now that Tesla is at $770B.
> What else am I missing?
Now filter for SoftBank.
I don't think it makes sense to include Juicero in the list. There is no way that whatever Ive and Altman are cooking up is going to be as obviously stupid as Juicero.
I dunno, I'd take a glass of juice over cryptocurrency from a retina-scanning orb.
At least the creepy orb is trying to solve a problem that actually exists. Namely how do you know the guy you are talking to is a real person? And especially, how do you know the guy you are donating money to is a real person?
A glass of juice solves the problem of wanting a glass of juice... Both products solve a problem in a worse way than existing solutions.
I’ll know he’s a real person if he gives me a glass of juice.
> There is no way that whatever Ive and Altman are cooking up is going to be as obviously stupid as Juicero.
Worldcoin is easily 10x more stupid than Juicero. A glass of overpriced juice is much more benign and far more useful than retina scams and cryptocurrency.
"There is no way" famous last words...
21 Inc, $ 116m: https://cointelegraph.com/news/bitcoin-startup-21-inc-reveal...
What I wonder is, how do you raise 100M before launch? Like, they can’t possibly have a 100M valuation. So do they just dilute like crazy or what?
(I guess Tesla was largely funded by Musk kn the beginning, but I doubt that’s the standard)
Firefly Space
I am sure it will be very thin….
I still remember when everybody suddenly added “COM” or “NET” to product and company names during the .COM bubble. Same with “AI” today. Good PR but ultimately meaningless.
They are designing the Niemann AI chess beads but for real this time.
On a tangent, was that ever remotely proven. Like even a smidgen of a hint of evidence?
No.
The interesting thing is they have the bankroll to actually make this work and it actually makes a lot of sense.
I remember watching a Google I/O presentation back in 2016 when voice assistants were trending. There was all sorts of ideas like APIs to access google assistant, voice/chat being the new interface etc.
The future presented there made so much intuitive sense to me and now the tech has finally caught up. The current static state off affairs has clearly outlived its usefulness.
Its not all sunshine and roses, Apple is an order of magnitude better position to take this leap then OpenAI. I know this may be a little out there but I think Apple made the right move by not jumping the gun with LLMs, IMO it makes more to let the dust settle and either partner with one of the winners or pick up the open standard (Llama 5?)
There was an Information article a few weeks ago on Apple increasing its daily spend to millions of dollars in ML training costs per day. With their historic desire for control over the whole stack, I'd put money on them rolling their own architecture.
Thats a start but if they were really going for it, it would look more like a multi-billion dollar investment with M&A, head hunting, training infra, partnerships etc
Atm its looks more like getting they lay of the land, which IMO is the smart move.
Apple has invested a lot in making sure that they STAY in "an order of magnitude better position". They know that the ultimate winner in personalized AIs will be whoever has the best edge hardware. That is why they have been investing so heavily in special purpose on-device chips for running neural networks.
If they take that leap. Apple was pretty early to the voice assistant game but they have been content to never actually do anything interesting with it. Siri is easily the worst mainstream product in its category, it has always baffled me how long they let this be the case.
Some of that makes sense. The three companies with the biggest voice assistants are Apple, Amazon, and Google. What are they all getting out of it? Google gets more data which is what they want. Amazon gets something, maybe knowing more about you, maybe some small margin of orders. Apple does things on-device so they get?
I'm not saying I like that Siri is terrible. I'm a Siri/Homepod user and somehow it seems worse each year, but possibly Apple sees little upside to Siri compared to adding a programmable button to the iPhone.
Something I've thought about for the past year, we can totally have real life PokeDexes now. It's all there; vision, voice, and reasoning. Just needs to be packaged with enough hardware to power local models, which smartphones aren't quite there yet. I want one.
The app is called seek and it works very well with local models.
I can’t help but wonder if Ive feels responsible for smartphone addiction. And if he hopes he might reshape our relationship to screens by introducing a new way to interact with technology. And if, at the end of the day, he might just create a new, deeper addiction.
Does this signal that OpenAI is seeing the limits of their own technology being hit just around the corner?
There is already no shortage of investment and hype around their core product - but it definitely does not seem like something you'd want to cram into a physical product, and definitely definitely not something you'd need $1B to develop.
This smells like "Open"AI is on-ramping to the final stage, cashout.
1. Raise $1B for a device blessed with Ive cool-factor for maximum hype and sales
2. Launch this device for $999 ($1299?)
3. Record-breaking IPO shortly after
4. ???
5. Everyone realizes they don't really need a $1k device to tell them "As an AI language model, I can't answer that question" fifteen times per day
A device that is 1000x better than Siri + Alexa + Cortona could be worth tens of billions of dollars a year in revenue. Revenue that OpenAi can use to fund UBI.
Someone will build this device, might as well be OpenAi (+ Jony Ive).
Wait, SoftBank might still have one billion after all their bad bets and failures?!
I'm at a complete loss of words... seriously at this point just seize the money and hand it out at random. That outcome is sure to be better and more productive than anything SB has done or will do
ARM, ByteDance, Compass etc. they’re making a killing even after all those losses.
Softbank again? Another WeWork?
Never mind, the device will be thin and will break in 2 weeks if you dare use it outside a temperature controlled clean room.
This collection of names does not inspire confidence, with all due respect to sir Jony for past accomplishments...
The answer to what they're building is written on Karpathy's twitter - "A kind of Jarvis"
This is a ploy to extract money from Masayoshi Son, the biggest fad chaser on the block.
When SoftBank is in you know it's time to get out.
[dupe]
More discussion yesterday: https://news.ycombinator.com/item?id=37681663