So Dramatic 📧
Beyond the (very important) matter of trust, I find Microsoft's latest stab at consumer AI with Copilot – not to be confused with their many other Copilots – to be pretty awkward. The product itself is fine and they're trying some interesting things. But are you really going to use it alongside ChatGPT? In particular, the voice version of ChatGPT? Microsoft, of course, will say 'sure!' But this particular space feels a lot more zero sum, I don't care how provocative or not the chatbot is. What does a certain $150B+ startup think about Copilot, which it's helping to build up to potentially tear ChatGPT down? Awkward.
I'm not sure that's more dramatic than the situation around the idea of Meta's Ray-Ban smart glasses being Terminator-style doxxing wearables. But it's close.
Some Thoughts...
📈 OpenAI Valuation Surpasses Every VC-backed IPO – Beyond being the largest VC-led private round ever raised at $6.6B, the $157B valuation pushes the company past any other American company at the time of going public. Assuming ChatGPT is correct – at these valuations, let's hope so! – Uber was around $83B. Coinbase $86B. Facebook $104B. I say "American" because Alibaba stood at $168B when it went public a decade ago. Still, that's well within reach. Saudi Aramco, maybe less so. $1.7 trillion at the time of their IPO in 2019. Pivot to oil? A few other fun ones just to note how times change. Google? $23B at time of IPO in 2004. Apple? $1.8B in 1980. NVIDIA? $626M – yes, million – in 1999. Microsoft? $350M in 1986. [Pitchbook]
😎 Meta Trains its AI on Images You Ask Ray-Ban Meta AI to Analyze – In other big-headline-but-little-surprise news around Meta's smart glasses, I mean of course they use images you ask their AI to look at to help train said AI. And guess what? So will any other company with similar products. If anything, this just showcases why the other tech giants probably feel like they need a competitor to these smart glasses – it's real-world training data galore! We'll see where Apple ultimately nets out here. Right now, they're positioning that they won't do this with Apple Intelligence, but there's some wiggle room here given the on-device and private cloud elements. [TechCrunch]
📡 DirecTV CEO Sees a Future Beyond Satellites in Big Pay-TV Merger – If I'm reading this right, it seems like post-merger, DirecTV/Dish aims to be the bundler of all streaming services. That sounds nice on paper – someone needs to unify all of this, through UI, if nothing else – but it's not entirely clear to me why a satellite television player thinks they can do this. I guess they think because when combined, they'll be the largest single player with some amount of leverage, but how much leverage is that really? They're also losing subscribers faster than anyone else. And what's the end-state here? That satellite TV is gone but DirecTV is the hub people use for all their streaming services? Again, maybe, but it's hard to believe they can pull that off? Netflix tends not to play ball on such things, will they here? [Bloomberg 🔒]
🤖 Amazon’s New Fire Tablets Have AI Inside – It's pretty table-stakes AI stuff (a writing assistant, a wallpaper generator, etc), but the bigger news may be that they've upped the RAM in the device in order to run the technology. One oddity: zero mention of what is powering the AI features. Is it Anthropic or Amazon's own technology – one of their many AI tools? Perhaps they're waiting to say more until they launch the updated – Remarkable? – Alexa, which should be any day now... [The Verge]
📺 Amazon to Increase Number of Advertisements on Prime Video — I don’t know about you, but few things say “Day Two” to me more than the notion of Amazon ramping up ads on Prime Video because they haven’t seen a revolt after the initial roll-out of said ads. That sure feels like the focus is more on the advertisers and the business than the customers. I guess you could argue that the ads allow them to keep prices down and content up, but that’s indirectly customer-focused at best as more advertising will certainly make the customer experience worse. But hey, the ads seem to be working, and the rest of the streaming industry — including, notably, YouTube — doesn’t seem particularly happy about the way Amazon has flooded the market here already. Get ready for the dam to open more! (I’m one of the apparently less than 20% that pay extra to not see the ads — as I can’t stand ads, and television ads in particular.) [Financial Times 🔒]
Some Analysis...
Some Links...
- J Allard... now that's a name I've not heard in a loooong time. Because he left Microsoft, where he led up Zune, Xbox, and the ill-fated Courier tablet, some 14 years ago. But now it seems that his old colleague, Panos Panay, has coaxed him back into the game at Amazon to work on "new ideas". [The Verge]
- A subtle, but seemingly important aside to the OpenAI fundraise: there will be a secondary offer for employees, but it will be meant for the company itself to buy back shares. Why? Undoubtedly to help them bolster the equity situation after the for-profit shift – which will involve converting profit-shareholders into equity shareholders. It's going to be complicated and messy. [The Information 🔒]
- One more thing: I've seen a number of stories – including this one – just parroting the idea that Microsoft gave OpenAI hundreds of millions of dollars (or more!) in this latest funding round, but why are we so sure it's not just in the form of credits once again? Given how much OpenAI needs to pay Microsoft for server/compute costs, isn't it just as good as cash without Microsoft needing to fork over actual cash? [Financial Times 🔒]
- Related: no surprise that Microsoft's data center real estate costs are rising fast – though some analysts are shocked at just how fast. At some point, Wall Street as a whole will turn against these expenses and that will undoubtedly be well before anyone sees any correlated rise in profits from said spend. [CNBC]
- Seemingly lost amidst all the funding hype, OpenAI also held a developer day a couple days ago. The key announcements seemed to be a few new developer features to match those of Anthropic, a simpler way to do model “distillation” to use larger models to create smaller ones tailored for specific tasks, and a real-time API for the service’s new voice capabilities. [The Information 🔒]
- Pete Wells, no longer just needing to review restaurants, goes on a hunt to figure out how chicken tenders conquered America. Hint: Burger King. [NYT]
A Golden Oldie...
And I Quote...
“We’ve been doing some work with the Meta Ray-Ban guys to see if there’s something interesting there."
-- Tom Hale, CEO of Oura, which just announced their 'Ring 4' line. As I noted a couple days ago on the word that this new smart ring was coming soon:
Given what Meta is doing with their new neural wristband for their "Orion" glasses, they're showing a way to use such wearables to make interactions more seamless, subtle, and accurate than simply relying on cameras. Apple should do this with the Watch – this report states they're not currently working on a ring, but have clearly thought about the form factor – and Oura should partner with someone on such functionality.
Well, I didn't expect them to partner with Meta on this, but it also makes some sense. As cool as the neural wristband may be (have I mentioned I was an investor in the technology that became this product back in the day? Oh, many times?), it's probably not something you're going to wear all the time, at least not yet. The ring though – it now has up to eight days of battery life on a charge.