investing all my time and money in food and fintech infrastructure at @NycaPartners @DirtyParrotVC
New York, NY
Joined August 2017
- Tweets646
- Following229
- Followers299
- Likes2.7K
Jas retweeted
Funding rounds aren't milestones. It shouldn't be your goal to raise a series A. It should be your goal not to need to.
Using DoorDash isn’t complicated and also DoorDash is for most people a browsing place to find inspiration and order. Much like shopping. It’s less common you know EXACTLY what you want and you quickly order it. Think outside your bubble of how most people order dinner on DoorDash.
You can’t really browse on a personal agent platform, you can ask for inspiration but that’s typically not how the moment of sparks and joy happen for consumers. Personal agents still have to work hard to capture your taste, they need time to learn and understand what you like and don’t like. The best consumer products give people something that works the moment they use it, that’s how you get addicted to using consumer products.
Otherwise, when shopping for something, people want options and visuals and swiping and pictures. Personal agents completely strip away the browsing experience which is why their use case needs rethinking.
DoorDash says they have only received "several hundred orders in total" from personal agents like Muse.
1
2
201
Jas retweeted
I’ve long believed getting to talk to a person will be a white glove, premium service and agents are only accelerating the shift to this
Darioooo….. we talked about this.
We discovered a sophisticated recording device that was hidden inside one of our offices. We used our models to understand the device and where it had come from, identifying that a specific employee had planted it a few days earlier. We referred the matter to the FBI, and we’re grateful to the agents and prosecutors who moved quickly to make an arrest. The FBI’s investigation is ongoing and we will continue to support their work.
At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data. The employee was terminated.
We take protecting our technology, data, and people seriously, and as a matter of national importance. Our security program is designed to protect against a wide range of threat actors, from insider risks to cybercriminals to nation states, and this case is an example of how AI can help address these sophisticated security threats.
2
79
Jas retweeted
Uber: We have many cars 1 minute away, see?
Me: I’ll take it.
Uber: Great! Give me 5 minutes to find a real one.
Me: What?
Uber: Locked! <map zooms out> He’ll be here in 12 minutes!
NOW THIS I LIKE. Finally hardware that every household might actually buy!!
I would definitely replace my google home with this. 👏👏
3
158
This is good. I genuinely believe good use cases for PAIs are savings, debt repayment/settlement, active budgeting, expense tracking, retirement planning, medical bill management, insurance premium negotiation, etc.
You don’t need verticalized apps to do this, but PAIs can and should be marketed to do all these tasks.
I get why you’d want to market access to luxury (flights, shopping, etc) because it’s aspirational. Aspirational always sells well with consumers, but the problem here is agents don’t help you achieve last mile of the aspiration, which is affording the aspiration, they just get you easier access to the aspiration.
Instead the flow can work end to end - agent helps you save for the flight, and then it is constantly matching your savings to best priced flight and once it finds the equilibrium, it books the flight. Or you save for the shoes you really want, and maybe they go on sale, boom agent buys them for you.
These moments of delight capture, engage and retain users, but also encourage good behavior and financial improvement for Americans that desperately need this. Imagine a dopamine hit from savings rather than gambling, that’s a world we can build with agents.
Jas retweeted
Cash / Dividend Businesses
Most AI data businesses should run as cash / dividend businesses. They can get profitable very quickly and start throwing off cash immediately.
(PS: This is another lesson from Ads days. Many SEM businesses twenty years ago started doing this after they realized it was easy to build cash revenues but hard to build durable equity value).
Implications:
- Compensation should be revenue share vs fixed salaries: Both employees and contractors should be on a revenue share model, where they get some % of the revenue they bring in. (There are various ways to attribute revenue). I've recently met a few companies that are doing this.
- Only raise a small angel round of sub-$1M: You can easily get Labs contracts on a small raise. Not going to VCs also gives you the flexibility to dividend out profits. VCs are not really equipped or structured to take dividends.
Rough math:
-- Raise $1M on $10M post. Investor owns 10%.
-- Get to $50M gross revenue run rate by end of year 1, $35M net revenue run rate, which means $17M in-year revenue. Say $10M profit (after lending costs due to working capital issues). Dividend out 80% while keeping 20% on balance sheet. That's $8m dividend-ed out in year 1. Investor gets 80% back!
Ping me if you're interested in finding an investor for this model (Small raise + Dividends) who can also help you with working capital stuff, which is becoming a bigger and bigger issue.
agent to agent resolution and negotiations. it's happening!!!!
@atulajoy and I wrote about this 6months ago in our commerce without clicks piece.
thecommercejournal.com/comme…
Native customer service in Instinct
Nobody likes talking to customer service. Instinct should handle those tasks for you.
With Instinct Concierge, we enabled Instinct to make calls on your behalf. Now, we’re taking this a step further. When your Instinct connects with participating businesses on @SierraPlatform and @DecagonAI, the agents will be able to work together to solve your problem with speed, reliability, and safety without making a phone call.
I’m excited to announce that Instinct will be partnering with Sierra and Decagon to bring customer service experiences directly to your personal agent.
Customer service tasks like:
- Troubleshooting your home wifi issues
- Finding out when a delayed package will arrive
- Figuring out why your bill is higher than last month
- Getting clarity on a confusing return or cancellation policy
This was my original problem with @TownAI . My inbox became total slop but I realized you need to shut off the notifications and tell it to archive everything. And then send you a bulleted list once a day of all the emails that matter and then just run through those. If there’s stuff to respond to tell it specifically, only draft those emails. Also forward things you need addressed specifically, don’t let it proactively do shit for you. This is where you see incidents of “oh instinct sent my personal balance to the company slack”. Use separate ai agents and accounts for personal and work you dingus. Also maybe don’t connect financial info to the agents that are run by 10 20years olds that have never built risk systems before.
Anyways, garbage in, garbage out. So don’t put garbage in and you won’t get garbage out. 😊
Agents are only as good as you set them up to be.
Personal AI agents are making me so efficient! Now, whenever I get an email, I get a notification from Gmail as well as a text message from Instinct saying I got an email, and a notification from Muse saying I got an email, and...
1
80
Jas retweeted
Bahahahahahhaah
SpaceX looks to raise $40bn to buy Nvidia chips in financing led by Apollo ft.trib.al/AXTvVlK
Jas retweeted
things that give normal ppl a dopamine hit:
adding things to cart. window shopping. buying something. waiting for a package. booking a flight. looking at hotels. imagining a vacation. planning an escape from everyday life. browsing restaurants & booking.
& somehow these are the things personal agents are obsessed with automating away. pretty interesting.
i like this part: agent demos should focus on the basics: tax filing, chasing down shipping returns, clearing out inboxes, and bill paying. every day stuff.
it’s always felt like OpenAI is too focused on winning San Francisco. meanwhile, their competitor Muse is sweeping the rest of the nation and sitting at the top of the app store. this demo is a perfect example of this.
every tech leader wants to show agents booking flights. but more than half of Americans don’t even fly once in a given year. so the primary agent example we keep showing simply doesn’t resonate with half the country.
agent demos should focus on the basics: tax filing, chasing down shipping returns, clearing out inboxes, and bill paying. every day stuff.
sf is very important. especially if you want to win developers. i understand that. but Meta is looking like geniuses focusing on everyone else. and everyone else is a pretty big market. it's the largest one.
Jas retweeted
Ben Affleck (Hollywood star & Artists Equity CEO) talks about how he fine-tunes open video models by unfreezing weights and trained only the last cinematic layer so a film crew can hit real production standards.
for context, Ben Affleck founded InterPositive in 2022, a 16-person AI shop for film post and Netflix bought it in March 2026 for $587 mn in cash.
He needed that model because public video models were trained on his peers' films, and he did not think that was a real business. So InterPositive raised money, shot its own dataset for 8 months on a controlled stage, and used it only as late-stage training.
Each new film then trains a private model on its own dailies, so the production keeps the footage and the learning. That is the product Netflix paid $587 million for.
----
From "Bloomberg Live" YouTube channel, (link in comment)
Jas retweeted
People saying Apple should buy Instinct, but this hits all the AI Bingo card elements —
1) Anthropomorphic AI — the agent is cute and gets sad in the ad when you don't use it enough.
2) Gamification — feels like an addictive Nintendo game where you're encouraged to level up.
3) Stripe Link for payments — duh, what everyone uses until they bring payments in house.
4) Free space
5) A2A Social networking — your pickle finds other pickles? Do they talk about their dreams about electric sheep, too?
6) A2A communications and coordination.
Introducing Pickle, a personal AI you raise. It gets things done for you and makes friends of its own.
Meet yours: pickle.com