100%
1/1 predictions confirmed
2 pending · 0 expired
23
companies · 89 data points
if you can help uh the smaller seller, the smaller DTC brand, >> dare I say, even the mom and pop um have a fulfillment system that is competitive with what Amazon has built for itself, then they can better compete with the incumbents.
if you're going to stay private essentially forever and i was just talking to um the ceo of hippocratic ai for our ai show the other day about this and he was like why would i ever go public private markets are super deep and i want to have all this control and not deal with the headaches of being public i guess fair enough but if you also then ban or preclude secondary activity to provide liquidity to your investors you end up essentially locking up people's capital what feels like forever or o
there is a gap in performance from the systems that say tesla ourselves have today to get to general purpose driverless
you can see the amazing tesla announcing a one and a half billion of revenue a year with this like clearly there's product market fit with with that kind of product
yes there's there's three ways to bring autonomy to market right you can build your own cars that's what tesla's doing but then you're limited to just your own brand
you'll be able to call it on the uber app soon
I think Anthropic has significantly passed OpenAI on a lot of things. And I think the reason is that they've been more focused. Like what's happened to OpenAI, to focus on them for a second, is that because they were first and they've had to raise so much money to build these models, Sam had to go out and sell that they were going to enter. They needed a $10 trillion TAM, which means you've got to be in every market, right? And so I think that's their problem. They've been trying to do a little
I would so much rather have us have that time to work to get things secure than to be catching up for three to five months until our companies were of a sufficient quality to match them.
I think we should consider this mythos model, this mythos preview to be essentially a cyber weapon and perhaps a cyber weapon of mass destruction.
I like what they're doing because they're preparing us for that day.
I think they're going to IPO this year.
I think Anthropic has significantly passed OpenAI on a lot of things. And I think the reason is that they've been more focused.
Anthropic has also put together a $100 million credit fund, essentially saying, here, use the model up to $100 million of compute, to essentially harden these systems.
There are the companies that are sufficiently important that Anthropic is letting them have access to Mythos, and that means they can be ahead on both defense and offense. And also, we're now seeing a world in which smaller companies are just stuck outside the glass looking in.
I think that Anthropic is doing the right thing here by saying, hey, we're not going to release something this dangerous, you know, from just off the cuff, but it does create a situation in which we now have a very much two-tier economy.
Anthropic is now very far out in front with its new model called Mythos.
I love getting a diversity of projects all part of BitTensor because it shows what the project can do as a whole.
BitTensor is a decentralized network that uses crypto incentives to reward individuals who contribute useful AI models, compute or results to task specific subnets.
i think it proves my first answer right which is the marketing worked
the first consumer app to truly bring OpenClaw into the mainstream will become a unicorn.
Lots of great new models out there, really cheap.
One thing that I found really frustrating is getting my personal chat GPT instance up on my and then going over to Anthropic and Claude and then
people have been Claude-shotted, like one-shotted. They are addicted. Everybody thinks this is the end of employment and everybody's just going to have six Mac Minis on their desk running stuff.
Claude Cowork going crazy at Davos. All the tech folks were talking about Cowork. That's playing with Cowork. Very impressive. Then I see Claudebot going crazy this weekend. Something's happening with Claude, Claudebot, and Cowork, which is leading to everybody on X buying Mac Minis.
June 2022, Zipline Cited had done 325,000 deliveries by August of last year, 1.6 million plus. The most recent number is 2 million plus.
Platform 1 is what you guys got started building, which is a longer range, lower weight, lower capacity drone that you've used across Africa to deliver medicines to great effect.
It's a three-year deal, one year of exclusivity, which means that I don't think another company can land a similar contract with Disney in the near term.
Disney will invest a billion dollars, as we said.
Incepting a Jedi story in your brain using one of these and you ask Neuralink or one of the other people doing this, like, is that an infringing? Well, yes, because Neuralink will become the new Disney world. And so in five years, Disney will do a deal with Neuralink that you will be able to experience any of the Star Wars movies.
in the wake of the Gemini 3 launch, Alphabet shares shot up by 5%. Now, why do we care about a 5% move for a company? Well, the company's worth $3.5 trillion, which means that 5%, Jason, is worth about $175 billion, which means that I think the market just repaid Alphabet for all of its AI work ever in a single day.
So if you are Zipline and you want to compete with DoorDash, which I believe is their business models to compete with DoorDash,
there aren't that many categories that like model labs are known to optimize for today and it's just going to blow up. So I think the, uh, I mean, I think it's, it's pretty amazing how, how much more work there, there continues, uh, to be.
I think it's absolutely insane how, uh, how like underutilized LLM still are. Um, there are, you know, I, I, I mean, I, on top of just running into a lot of repetitive tasks that even engineers continue to do today. Um, like the, the, I think the most exciting sort of glimpse into how many more tasks can be automated is the, the new, like types of training data that the model labs want.
the last component is that there's a whole bunch of inference adjacent software that people really want. Um, things like observability, web search, um, like improves team management, the ability to like limit certain parts of your team programmatically, um, easy, uh, like multimodal support, um, PDF parsing and PDF management, file context management, all this stuff that like, it kind of like, it kind of becomes really powerful if you have it managed very closely with your inference. Um, those t
the other way we do it is by doing tons of volume on providers and, um, and just like working with providers to, uh, like, you know, do volume discounts. And then we often give those to users too.
So it's there's, there's three ways there there's this credit top up fee, which, um, which is it's basically for off the shelf inference. If you want to do volume with us, you can contact us through our enterprise page and, uh, get that fee down quite a bit.
Models are like content. Every time there's a new model, it's like, it's like a holiday. You know, there's, there's, uh, um, you know, there's a whole bunch of problems that might get solved. It's a new marketing moment for all the apps that suddenly can support the new model. It's, uh, it's like a new person being born. And, um, and, and, and when that happens, people develop these like personal relationships with the model. And then they, they like understand that, and then they want to go to
There's this, uh, like LLM power user community that, um, that we really identify ourselves with. That's kind of how the company got started.
now the model providers and the model labs like see the value in routers. And it's now this kind of new AI product category, like matters, routers are around to stay.
from the provider's point of view, you can think of us kind of like a distributor, where we're just helping you discover new users.
a lot of the people who try out new models when they launched would probably never have done it had it not been really easy to just switch and explore it.
by making it easier to sample, to switch and so forth, you also effectively increase the quantum of competition inside of the AI market.
OpenRouter gives you a single pane of glass for discovering new models, discovering existing ones that you have access to, doing inference the exact same way through all of them. So we normalize and standardize and fix errors with all requests. And we upgrade the quality of inference pretty significantly.
You have heard of this company. It's worth a lot of money. It's growing very quickly.
Just to give people a little bit of context here, it is a San Francisco-based startup with roots in Kazakhstan. And they offer to consumers, essentially, a lot of AI models you can play with. Some are open source, some are not. And you pay for a monthly number of credits, essentially.
The reason why this is blowing up, and I wanted to show it to everybody, is that the face swap is bonkers good.
so then therefore hux really is uh something you take with you as you drive as you cook as you eat so it's designed to be an additive to your daily life so it can fit into time that you already have you don't have to make room for it
on the habit point you guys had to make a choice between text and audio and you made the decision to have hux talk to you and not only in one voice but one thing i noticed was there was a back and forth between a male voice and a female voice on my end and that did make it pretty darn engaging
you made a product that everyone knows i've actually like no joke notebook lm is in our circles a smash hit so i really do feel like you could you know brush a little dust off and be proud of it
hux came out um in september out of kind of early access or beta
when i signed up for hux it said hey can you connect your calendar can you connect your email and then what are your major interests and then from there it started to essentially tell me about things that i'm interested in we're talking about apple and stock market and startups and i was like okay well fair enough that's pretty much dead on
i love playing with new apps and i got to play with hux and frankly i really like it
magic school seems to be doing quite well it raised a 45 million series b earlier this year
it's a really cool idea it's growing like a weed
And what a great real estate play. If you wanted to build a bunch of batteries and put them somewhere, you've got to buy some land here. You just, you know, get a little 10 by 10 spot on people's existing properties and you get to distribute all this power. It's just super brilliant.
These are 800 bucks from Meta and up for the bracelet and glasses.
one thing we're seeing Waymo just do now is go to D.C., go to Denver, go to places like Seattle where it rains and where it snows.
If you know Uber, you know, they used to have a self-driving arm that eventually got sold off, but now they are back partnering with Lucid and its gravity SUV and EV and Neuro's self-driving technology.
Databricks is one of the most valuable startups in the world, period.
the biggest financial news this morning in the technology world jason is that core weave is going to buy core scientific core we have of course was an ipo earlier this year we talked about on the show a kind of ad nauseum
we have also seen news that Apple might work with either Anthropic or OpenAI to source their models as opposed to what it's built in-house. And if they do that, the need for perplexity goes down a little bit, Lon. But I'll also just say Apple has a history of small acquisitions, not big ones. And that's also leading against this being a possibility.
Meta's enormous AI push is now really impacting the startup world. Recall Jason, that they are trying to build a new super intelligence team. This is where the scale AI quasi acquisition came into play.
Sam Altman, who we all know from his days at YC and currently runs OpenAI for all intents and purposes, said that the company now has enough going on across three different things, its research arm, its build out of mass compute and also its applications business. And he's going to hand the application site over to Fiji.
my read of this is that it's not really a diss against Instacart, but that it's probably a lot more exciting to go run a business that is growing at one bajillion percent a year at the absolute forefront of technology than it is to try to grind out another five percent growth in your grocery delivery company.
NVIDIA has been dethroned overnight by DeepSeq and is now the third most valuable company compared to the first.
if there's any group of people out there who could put together this kind of capital for a project, Sam Allman's famous for his dealmaking, Masa's son and SoftBank have done some of the biggest investment vehicles in the history of investment, as far as I can recall.
The gist is that there's going to be a new company called Project Stargate that will start with $100 billion to build out infra that's going to support OpenAI and its development.
Microsoft will no longer now be the exclusive cloud provider of OpenAI, but they will retain right of first refusal, and they have longstanding commercial agreements with OpenAI. That means that Microsoft OpenAI partnership is not broken, but it's certainly a little bit diminished now that this new product has come online.
OpenAI is doing equity funding, and they're the lead partner on operational responsibility for the project, and they are, quote, a key initial technology provider.
Now, you could argue that Google got worse over time, but essentially everyone was still chasing that Google love.
And so Apple might be time for another act now that the Division Pro has yet to catalyze revenue growth.
However, more recently, iPhone sales, much like smartphone sales around the world, have slowed, you might even say stagnated.
I don't think Meta is in any danger of running out of money. They can afford to both invest in AI and reality labs to the tune of several billion dollars a year, but at the same time, they'd like to make more money. And if Meta was able to double its profit, it'd be worth a lot more.
with rising search query load, an AI-first approach that is, I think, cleaner than what Google offers today, a subscription business, and a nascent ads business, well, you can kind of understand why VCs are excited about the company.
I view it as kind of a binary bet. If you value Perplexity on its fundamentals today, it's probably a little bit expensive at the $9 billion mark. But if you think it's going to get 1, 2, 3, 5, 10% of the search market down the road, then it is absolutely on sale today.
SEMrush reports that Google handles about 5.9 million searches a minute, so Perplexity is still orders of magnitude smaller than the search incumbent. But here's the thing. I don't think that really matters. What matters is that search is incredibly valuable. And that means that every point of search market share that a company can control will generate an enormous amount of wealth.
after handling around 500 million queries in 2023, an impressive number, Perplexity served $250 million in July alone. And more recently, in late October, the startup said that it had scaled its query load to $100 million per week, or about $400 million per month.
so i guess the way that i'm going to approach this jason after going through my settings today and tinkering with it is that for all my messaging apps i'm just going to have apple intelligence turned off period but i don't actually care if apple intelligence knows what i use in google maps because it might help me get places faster so i think that's where i'm willing to give up a little bit of my privacy ethos if you will for convenience but if you're going to use signal which is as you said ind
apple has made such a push though about being the privacy first provider for mobile operating systems and hardware for so long that they should have pretty good consumer trust here but i don't think that's enough to make us feel good
i think the media has been very positive about apple's work here but i don't think that general media privacy educational awareness is particularly high and so i think apple's getting a lot of credit from the media for saying hey we're going to take a look at this we're going to be serious about it i don't think that's convinced the actual privacy advocates
going through privacy-minded forums talking to people about this my impression is that apple's setup is the gold standard for this type of technology
a lot of people think that apple has offered a reasonable set of privacy guidelines for their use of ai which is on-screen learning and so forth they're on-device learning but people are still very concerned that some ai processing if you will some of this learning that apple intelligence wants to do will be done off device
i think the concern that people had here is that apple was turning this on automatically for signal if you turned on apple intelligence and i think that's a mistake from a design perspective
last week hubspot announced that it's going to purchase hash flow a startup that had raised 10 figures while private
if you recall the Google glass days they had the little translation app and so you could look at a street sign and it would auto translate it for you I got to play with that blew my mind
Microsoft is currently working to actually get a nuclear power plant to turn back on just so it can help power its data centers with low carbon energy.
Google is one of those companies that has its own vision and a long time to grow into it. And it has its own way of doing things. And I guarantee you it's not the way whiz does things because whiz is small and Google is enormous.
Google was offering 23 billion for a 350 million ARR company, which is a pretty rich multiple for the current time.
right now, Nvidia is worth more than Apple. And so what we're discussing here is where the value will accrue. Will it be in the people making the chips that power the AI or in the companies that are then using the AI to take it to the end user, the consumers in this case. So it's kind of a fun race and challenge to see what's going to come out on top.