Magnus (@cmagnust) header
Magnus (@cmagnust) profile picture

Re-thinking human machine interaction @domeralabs

NYC / Zurich
Joined September 2017
Photo by Magnus (@cmagnust)Photo by Magnus (@cmagnust)Photo by Magnus (@cmagnust)Photo by Magnus (@cmagnust)Photo by Magnus (@cmagnust)Photo by Magnus (@cmagnust)Photo by Magnus (@cmagnust)Photo by Magnus (@cmagnust)
Magnus retweeted
crazy that these are 'side' stats that didnt even get an image to add more context, standard reserve's eth/usdg pool is not just the 2nd largest eth/usdg pool by vol, it is the 2nd largest pool in general on robinhood chain crazy what concentrated liq and v4 hooks can do innit
Replying to @standard_rsv
Some additional data points: The Reserve has near total ownership of SKHY trading at 99%. We seeded an ETH/USDG pool to prove our liquidity efficiency and it is now the 2nd most traded ETH pool on all of Robinhood Chain. The average LP APY across pools was 58%.
5
10
82
4,403
Apathy, seems part of the equation while I also strongly believe that lacking product discovery is another one. AI companies are failing to educate the non tech/vc retail normies about its capabilities and where it can be useful in everyday life.
The winners in consumer AI will be whoever solves the highlighted sections on this chart 👇 Privacy and trust concerns usually fade as people find products that add value to their lives The primary battle right now is against apathy
2
5
159
Technical moats are mostly over at this point. If you’re building something that is meant to be durable, the only moats now are: • Proprietary data • Network effects • A physical component
1,145
2,082
907
25,720
3,342,403
Fragmentation is bad UX, and when general consensus is that there will be >1, the way humans interact with these has to change. A true AI-native OS (voice-based) will be the platform for which the hardware does not yet exist. but soon
Been asked this question many times in the past few days. Curious for others’ views: How many different consumer (non-work) agents do you think people will be using in 3 years:
36%One for everything
48%A few specific ones
14%As many as I have apps
3%As many as I have friends
634 votes • Final results
6
212
The amount of use cases for personal AI assistants outside of the vc/tech bubble is insane.
2
8
429
Huge congrats to the @hume_ai. After trying out the Expression Measurement API I am fascinated. Really excited to show what this enables beyond obvious use-cases, eg customers service and how this can solve one of current models' biggest challenges: long context windows
Someone says "I'm fine." You know they aren't. You heard it - the pause, the flat tone. Voice AI doesn't. It all gets dropped at transcription. So we built the Expression Measurement API. 48+ emotions, 600+ descriptors, 50+ languages, audio and video, real-time. Conversation was always more than words.
3
5
172
app interface > AI assistants & apps > plugins/connectors Next: screen based interaction > voice based interaction & screen based hardware > voice based hardware Owning the AI native hardware + OS while allowing special purpose assitants to plugin vertically is the holy grail
re: consumer AI travel, shopping, restaurant reservations are the boring trio of consumer AI now. anyone big enough can do it easy. especially with meta's infinity partnership spree, hard to top them. i'm rather bullish on the ones who are trying to eradicate friction in one nuanced consumer use case at a time, rather than those trying the jaded pitch of "AI used to tell you how to do stuff, but now it can go and do it for you." it's like selling a less known cheap taco next to taco bell.
2
4
140
Deeply knowing your user and building a long-term relationship is the moat. > How to store / classify information and facts about the user AI models are not sticky because of that, everyone is jumping from model release to model release.
If every AI agent has access to the same supply, where does the moat come from? Personal AI agents are becoming marketplaces, but the supply is mostly the same for everyone. That’s what makes Instinct creating its own supply so telling. If partners have no reason to give you exclusivity, you have to manufacture it yourself. To me, that’s less a moat and more proof of how little defensibility exists in the core personal AI model today.
1
4
176
The final form factor has not arrived yet but it's obious where it's heading: voice-based hardware + AI native OS including a main bot + vertical integrations of purpose-built special areas
I think fundamentally we've ~now arrived at the correct form factor for mass consumer AI: It was never going to be Google's search box but with AI, and a left-hand list of Chats was a 18mo experiment that has reached 100M users but won't be the thing that will scale to 1B users for next 10yrs — it will be messaging with a single bot, occasionally split to specific bots for special areas of work/life. Still ton of innovation to be done, e.g., in the bots–bots collab mechanisms, credentials, computer use, and generally optimizing the heck out of everything … … but the form factor has settled; the iPhone Classic has landed; the News Feed has been established.
2
152
The fact that 90% of Instinct users using voice for input is absolutely mindblowing. Apps become plugins, chat becomes voice and the end goal of this is AI native (voice-native) hardware.
My conversation with Noah Shinn (@noahrshinn), founder of Instinct. Noah is building a personal AI assistant. It's still invite only, has spent nothing on marketing, and is growing roughly 10% A DAY. This is his first long conversation about the company. We discuss: - Why Instinct doesn't have an app - Buying compute months ahead of exponential demand - How users learn to trust it with a credit card - Safety and security - Agents coordinating with other people's agents - Instinct's business model - Apps built on consumer inertia - and more Enjoy! Timestamps: 0:00 Intro 4:11 What people are using AI agents for 15:07 Rethinking travel, reservations, and the internet 22:43 Trust, privacy, and personal data 27:50 The business model behind Instinct 38:04 How existing businesses will adapt 47:55 Designing a personal assistant people love 53:15 Growth, compute, and competing with Big Tech 1:11:44 What’s next for Instinct and personal AI
5
712
My conversation with Noah Shinn (@noahrshinn), founder of Instinct. Noah is building a personal AI assistant. It's still invite only, has spent nothing on marketing, and is growing roughly 10% A DAY. This is his first long conversation about the company. We discuss: - Why Instinct doesn't have an app - Buying compute months ahead of exponential demand - How users learn to trust it with a credit card - Safety and security - Agents coordinating with other people's agents - Instinct's business model - Apps built on consumer inertia - and more Enjoy! Timestamps: 0:00 Intro 4:11 What people are using AI agents for 15:07 Rethinking travel, reservations, and the internet 22:43 Trust, privacy, and personal data 27:50 The business model behind Instinct 38:04 How existing businesses will adapt 47:55 Designing a personal assistant people love 53:15 Growth, compute, and competing with Big Tech 1:11:44 What’s next for Instinct and personal AI
149
226
213
2,388
1,832,179
Unless your app is for entertainment pivot hard
2
154
Big kudos to media company Axel Springer and its CEO, Mathias Döpfner, for honoring Peter Thiel - the most successful German ever. This is a long-overdue recognition for Peter in Germany. The entire award evening - including @bariweiss’s amazing award speech, as well as the podcast Peter did with Matthias before - should be required viewing for every German politician, journalist, and, frankly, everyone. The world is changing at record speed, while Germany is falling asleep at the wheel. It’s time to wake up. welt.de/wirtschaft/video6ab6…
16
16
2
259
13,768
Owning the OS allows for a major UX gains and better knowing the user. Hardware is hard but hot
1
1
287
rip if you are building an app
1
3
175
We are here: app interface > AI assistants apps > plugins/connectors This is where we are heading: screen based interaction > voice based interaction screen based hardware > voice based hardware
Replying to @finkd
And we've been building a joyful little device for talking to your Muse that fits on your keychain. Muse Charm -- shipping in December.
2
242
A parallel race will happen on the hardware side. When you have a proactive voice-first assistant, what do you need a screen for? Tremendous hardware shift incoming
i understand why a sharp investor would buy at 10b, but not for me to fight to then attempt to land a 30-50B+ acquisition offer and very slim chance of actually winning the category. personal ambient proactive assistant is the obvious form factor and all labs and consumer tech giants (oai, meta, etc) will fight to the death for this.
1
1
169
Magnus retweeted
i understand why a sharp investor would buy at 10b, but not for me to fight to then attempt to land a 30-50B+ acquisition offer and very slim chance of actually winning the category. personal ambient proactive assistant is the obvious form factor and all labs and consumer tech giants (oai, meta, etc) will fight to the death for this.
Instinct looking to raise $1B, potentially at a $10B valuation. Compute costs are high and Noah doesn’t want to charge users for the product. Instinct has raised $350M to date, mostly uses open source models and wants to own chips and data centers theinformation.com/articles/…
4
3
3
132
26,665