Excited to announce @a16z is leading @probookai's Series A! (@arampell, @dhaber, @seema_amble)
Probook is building the AI front office for the trades, starting with dispatch.
They already run in hundreds of locations, from independents to PE-backed platforms 👇
A voice-to-voice model that is reliable and has function calling has also been a bit of a white whale in enterprise AI
Will be interesting to see if GPT-Live finally cracks that open 👀
This is a big, big step up for consumer voice
Most people don’t even know ChatGPT has a real-time voice function
And for those who have tried it, naturalness/reliability hasn’t quite been there yet to foster heavy usage
Excited to see a new wave of voice-pilled users here!
Introducing GPT-Live, a new generation of voice models for natural human-AI interaction.
Rolling out in ChatGPT starting today.
You’ll want to turn the sound on for this one.
In five years, early consumer AI interfaces are going to look so ancient
Users having to choose between models, modes, or types of work is holding back deeper adoption
Claude merging Cowork into Chat and ChatGPT merging the app, Codex, and Atlas into one are the first signs
New today as Claude Cowork starts rolling out to web and mobile: Chat and Cowork now share one home tab on web and desktop — one sidebar, one search, and one place for your Projects & Artifacts. Keep an eye out for even better Chat + Cowork integration soon, too.
This is kind of fascinating as it feels like the opposite of the J-space in some ways (things not written anywhere that it remembers)
It makes me wonder whether Claude tries to make itself forget things that it deems ultimately unhelpful even if relevant for the user
I often find in Claude’s reasoning traces it will have thoughts / opinions it does not say
…but then it denies having those thoughts until presented with screenshot evidence
It’s like Claude is either: (1) lying; or (2) not able to remember what it doesn’t say
New Anthropic research: A global workspace in language models.
Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.
We found a strikingly similar divide inside Claude.
Also have to give Google props for the “Ask anything” feature where you can follow up on a search result
It’s difficult to break the behavior of Googling something first
This feature makes it more likely you keep searches on Google vs running them elsewhere
Also, as good as Comet, Dia, Atlas, etc were - when they came out, computer use was slow and unreliable
The best agents now (OpenClaw, Manus, etc) seem to have a different architecture
This will either come to a browser-like experience, or ChatGPT etc will become the browser
I was an early and enthusiastic adopter of AI browsers
I don’t think the category is dead…but because switching costs are quite high, you need a 100x killer feature
Once agents in ChatGPT, Claude, etc got good at browser use - the first value prop of AI browsers was eroded
bro what happened to all those agentic browsers,
one day they were gonna "autonomously browse the web, book flights, do research", where are they now????
Beyond “Google Maps with natural language search, personalized to you” I think there’s a ton of interesting stuff they could do here
As agents start to get better (and people adopt them for more use cases), it’s a valuable data point in mapping where you spend time -> outcomes
PSA - ChatGPT (quietly) rolled out precise location sharing over the last few months
You can turn it on in Settings -> Data Control
Once enabled, you can get much more detailed recs automatically (ex. “vegetarian food within a five minute walk”)
Info from @nicoleperlroth’s fantastic podcast To Catch a Thief (season two!)
An entire network of thousands of North Korean spies assembled on Discord and was sharing best practices on how to get hired (and not fired)
One candidate submitted 26,000 applications to get 19 offers
TIL remote North Korean IT workers are infiltrating American companies to the extent that a common interview Q is
“Repeat after me: Kim Jong Un is an ugly fat pig”
If the candidate freezes and then disappears, they’ve been caught
Jodie Foster says Apple’s “F1” seemed like it “was made by AI” and written as if a computer was following “the structure that you would learn in school”:
“I don’t say this disparagingly — how could I? This movie went on to make millions of dollars. But I look at a movie like ‘F1’ and I’m like, ‘F1’ was made by AI. Wasn’t it? I mean, the structure was exactly the structure that you would learn in school. The actors say the lines exactly the way it would be written if a computer was writing exactly what would be the right thing for that time. And they were able to dominate the technology to make something big and beautiful and potentially where a lot of the information comes from other places.”
https://t.co/LZw8DKLlUb
I ran the numbers on this a few years back and it’s absolutely true 👇
Consumer is the ONE category where being a repeat founder is negatively correlated with success
The biggest reason (IMO) is you can’t execute your way into a working product - it requires a lot of luck/magic
.@markpinc: "The reality in consumer is, where you see traction, there's a good idea. Where you don't see traction, there's not a good idea."
"That might sound dumb. But in consumer you're better off backing an unproven entrepreneur who's found product market fit than a proven entrepreneur with no product market fit."