VP @ @arkham Founder of @indexifyapp
- Tweets 1,923
- Following 163
- Followers 446
- Likes 1,679
Pinned Tweet
Hark vs Instinct: Dinner analysis
@adcock_brett s Hark took 20 minutes to order me a pizza and got it wrong. Two pizzas, wrong restaurant.
@noahrshinn s Instinct emailed 50 restaurants to organize a private dinner for 20 people and got it right.
That’s backwards.
Finding a restaurant requires research and reasoning. But once you’ve identified the restaurant and what you want, ordering should be deterministic. Select the items, confirm the details, complete checkout. This is a 5 min process at most.
Finding a private dinner for 20 is the opposite. Open-ended research, dozens of venues, inconsistent replies, contracts, and fine print.
Instinct researched 50 venues, emailed each one, followed up on unanswered requests, and organized 41 responses into one comparison sheet.
It caught a contract quoting 20 guests but charging food for 15. It flagged “private” rooms that were actually curtains or mezzanines. It asked for approval before sending each batch.
The simple task failed. The complex one worked.
Agents are becoming remarkably good at open-ended research and communication, but still struggle with routine execution.
Once the reasoning is done, execution shouldn’t require more reasoning. It should follow defined, permissioned workflows.
That’s an infrastructure problem, not just a model intelligence problem.
6
2
29
6,055
Sebastian Gallic retweeted
Introducing Freeport Partners, an invite-only program for creators on X, Telegram, and Discord.
forms.gle/wd6sAatMWkmgESi88
We’ve had creators and partners reach out about working with Freeport. Today, we’re opening a program to make that happen.
Whether you create trading content, write market research, or run a community, we’d love to hear from you.
What you earn for each referred user:
• 50% of their trading fees, up to $250 per person.
• $10 when they deposit at least $100 and make one trade, paid after a 30-day hold.
• An additional $15 when they reach $1M in cumulative trading volume. No deadline.
Both bonuses are on top of existing fee sharing. Partners also get priority for future reward programs.
Apply below, it takes about 3 minutes.
We manually review every application and onboard accepted partners in batches announced here at @freeportmarkets.
9
6
28
2,271
Sebastian Gallic retweeted
personal agent demos should be focused on getting rid of activities that people dread and put off or that require a lot of cognitive effort, not pizza and flights.
they should one shot license renewals, DMV requests, insurance policy shopping, car trade ins, etc
56
16
393
13,099
Sebastian Gallic retweeted
X really is the best place to reach the right audience. Tons of new signups for Loci. The future of AI is private.
The greatest trick OpenAI ever pulled was convincing the world that sacrificing your privacy is the price of using AI.
We think that's insane.
Introducing Loci: your own private AI, inside the apps you already use.
→ text it from iMessage, WhatsApp, Telegram, Discord and Slack
→ runs 100% on your own computer
→ chats never leave your machine
Try it free: askloci.ai/chat
4
3
22
1,439
Sebastian Gallic retweeted
Oh you let me into your front door for dinner?
What do you mean I can't go through every drawer in the house and read your diary??
You let me in the front door???
Fucking moron.
3
1
11
227
AI personal agents have an access problem.
Connecting your inbox to complete one unrelated task shouldn’t give an agent free rein to analyze your relationships, unrelated emails, or anything else it finds.
You authorized a task. Not your entire digital life.
What’s missing is a permissions layer between the agent and the system of record that enforces why data can be accessed, not just whether it can.
The agent should only retrieve and retain what’s needed for that task. Anything more requires explicit approval.
And enforcement has to live in the platform, not in a prompt asking the agent to behave.
Access isn’t consent to every use.
If you’re building this layer, I’d love to compare notes. DM me.
1
5
1,426
This app is insanely farming your entire life when you order this pizza.
I connected my Gmail to order a pizza, and afterward the AI casually admitted it had collected information about my family, closest friends, businesses, travel habits, loyalty programs, favorite restaurants, and even Citi Bike usage.
How fucked up is it that an AI agent tasked with ordering a pizza is digging through your entire personal life?
This is a massive privacy problem with AI agents.
We're giving away $1,000,000 of free pizza right now
Download and order from the Hark app: hark.com/download
Type "order a free pizza" in chat. It's first-come, first-served, delivered only in the U.S. (review details below)
204
120
2,054
237,091
Sebastian Gallic retweeted
Interesting that this is almost exactly what we have at Arkham? I don't see Arkham as a reference.
more than half of all BTC supply is sitting in addresses with no known attribution
despite all the institutionalization and centralized ownership over the years, still a ton of bitcoin (maybe even most) is held by individuals in self-custody 🔐
4
3
28
4,962
You know what’s interesting, I feel like longing consultant companies (ie accenture) is actually the move
Counterintuitive but bc people are having a hard time identifying ROI on AI, they’re gonna get hired to deploy
208
Although I don’t totally disagree
Debt / Revenue is not a metric that defines credit worthiness - literally non-sensical.
CoreWeave $CRWV is not going bankrupt anytime soon, in my view.
Its weighted average interest rate has fallen from nearly 15% to 7.8%, while its latest convertible debt raise came at just 2.875%.
This allows CoreWeave to borrow significantly more capital without a proportional increase in interest expense.
And don't forget, Nvidia $NVDA owns roughly 9% of the company.
124
imagine conflating interest with yield
$CRWV isn’t paying 15% interest. Only 4% of their debt costs that.
I went through the Q2 10-Q line by line. Out of ~$35B in debt, exactly one loan sits at 15%. It’s $1.3B, due March 2028, and already being paid down.
The 15% chart going around is the market yield on their 2032 bond. A higher yield doesn’t change what CoreWeave pays. It works like a mortgage. If you locked in at 3% and rates go to 7%, you’re still paying 3%. CoreWeave locked that bond at 9.625% and still pays 9.625%.
What they actually pay
> $6.6B of convertibles at ~2%
> ~$10B of bonds at 9-10%
> Most secured loans around 9%
> ~8.4% blended across everything
They have roughly $20B of 2026 capex left to fund, and 2027 will need more. Their main funding tool has been loans backed by customer contracts. The investment grade one backed by Meta priced around 5.9% in March. If they keep borrowing that way, I think the math holds.
Zooming out, the AI trade still looks intact to me. $ASML reported record Q3 profit this morning and said lithography demand will outrun their supply in 2027 and 2028, even after adding ~30% capacity each year. Customers are already asking about 2029 and 2030 slots.
Demand is still outpacing supply everywhere I look. $CRWV $ASML
1
800
They’re rated bb- and rates are through the roof
It’s not a near term issue, it’s a medium term issue on the inability to forecast commute prices and underwrite the risk
Every bond of $CRWV is failing today. Things like this almost never happen unless there are struggles to make payments. These bonds are spread across multiple institutions. All are being dumped. 14%-15% yields
$ORCL $NBIS $IREN
68