Danil

580 posts

Danil banner
Danil

Danil

@danilk

First hire @withshepherd. Operator lessons from the trenches https://t.co/J3h2fVPJ6z My better half @ninaserr

San Francisco, CA Katılım Ocak 2020
158 Takip Edilen458 Takipçiler
Pratyush
Pratyush@pratyushbuddiga·
I order from sidewalk juice on @DoorDash for our office lunch, but starting yesterday, it seems like they started AI-fying the descriptions. The names are off, protein is now spelled wrong, random exclamation points, and even the descriptions are incorrect. I wonder if it’s an internal AI thing they’re testing or just on the vendor side? Gut is vendor implemented some startups janky marketing offering but pretty funny to see.
Pratyush tweet mediaPratyush tweet media
English
5
0
11
4.4K
Danil
Danil@danilk·
@elvissun Curious if you considered deploying Zoe setup into the cloud instead of being local? That would solve the memory issue permanently. What made you stick with localhost?
English
0
0
0
122
Elvis
Elvis@elvissun·
mac dilemma update (73 comments later): going with M4 Max 128GB. now. here's why: 1. opportunity cost - every day my agents are bottlenecked is a day I'm not shipping (wwdc in june is 5 AI years away) 2. pre-AI pricing won't last - apple priced these before AI agent demand 3. vertical scaling = focus - managing network I/O across multiple minis is time I'd rather spend building 4. local = control - no rate limits, no API outages, full autonomy 5. memory > compute - RAM is the bottleneck for agent swarms, not CPU my plan now is to use the studio to power my agent swarm but rely on cloud models (codex 5.3 mostly), but one comment that stuck with me from @HEXtheantidote: "running a local model is the only way for them to achieve true autonomy" local models aren't there yet. but in 12 months? having a self-hosted AI might be as normal as having a router. Mac is the body. agent is the soul.
Elvis tweet media
Elvis@elvissun

mac dilemma: my mac mini 16gb can no longer handle the agent swarm options: - M4 max 128GB - $3.5K now - M3 ultra 96GB - $4k now - M4 ultra 128GB - $5k? at WWDC (june) would you buy now or wait?

English
35
5
174
72.4K
Danil
Danil@danilk·
4 years ago, before AI was everywhere, in the early days of Shepherd, @MoFromYYZ and I had to make a one-way door decision: buy a PAS or build it ourselves. Almost every insurtech founder we spoke to said the same thing: DON"T BUILD. We demoed nearly every major vendor. The price to value ratio wasn't there. So we built. That original PAS is now Shepherd’s internal operating system. Underwriting, pricing, workflows, risk signals all live in one core system. Most legacy carriers are still juggling 10+ systems of record. Our entire company works from one. I’m writing about the decision and what it unlocked in detail here: shepherdinsurance.com/blog/building-…
English
0
0
2
96
Danil
Danil@danilk·
The SWE/PM bottleneck is dying in front of our eyes ☠️ Old model: Idea → PRD → Mock → Implement → Ship. Now the constraint isn’t coordination. It’s judgment. I wrote more about this shift in my new blog post: danilk.substack.com/p/the-death-of…
Danil tweet media
English
0
1
2
50
Danil
Danil@danilk·
excellent piece on what it's like to deploy production-grade AI in the insurance domain recommend reading if you want to know how to bridge the gap between AI demos/eval benchmarks and outcomes for your customers shepherdinsurance.com/blog/your-eval…
English
0
0
0
32
Danil
Danil@danilk·
@RajaGhawi Excellent piece - excited for more
English
0
0
1
27
Danil
Danil@danilk·
LLM training and inference needs are skyrocketing — and so is energy demand across the United States. Proud to share that Shepherd will be helping insure the next wave of energy infrastructure powering our Great Country 🇺🇸 shepherdinsurance.com/blog/shepherd-…
English
0
0
1
63
Danil
Danil@danilk·
@AstasiaMyers If the cost of building software (AI Agents, systems of record) goes to 0 over the next X years, at what point would an enterprise be better served building these in-house vs having an AI budget at all?
English
0
0
0
462
Astasia Myers
Astasia Myers@AstasiaMyers·
Great piece: AI budgets are MASSIVE, and the core product moat becomes “agent context.” The strongest moats are: • depth and cleanliness of data • complexity and formalization of workflows • number of system integrations • embedded human-in-the-loop checkpoints
Aaron Levie@levie

x.com/i/article/2013…

English
25
30
511
122.7K
Danil retweetledi
Pratyush
Pratyush@pratyushbuddiga·
Great read on building an AI-services company. I work with two (@hanoverpark and @RivetTax) and Danil's lessons apply for anyone looking to build a generational company with this time of business model.
Danil@danilk

Just launched my blog: The Power of And. danilk.substack.com/p/the-power-of… It is about building AI-native services in the real world where software meets regulation, capital, compliance, and operations. First post: why cross-functional excellence is the real AI moat. And you can read a rant version of it in the X article below. If you are building AI that touches the spreadsheet economy (legal, consulting, insurance), this is for you.

English
1
3
73
24.1K
Danil
Danil@danilk·
Very bullish on this, excited to see more folks building in the space. Insurance should be an efficient market, customers should not be paying 30%+ for procuring an insurance policy. Can you imagine paying 1/3 of every transaction to Amex/Visa when swiping a card?
JD Ross@justindross

My last company, Opendoor ($7B), replaced real estate brokers. Today, my new company WithCoverage raised $42M to replace insurance brokers. It was led by Sequoia & Khosla, the first time @RoelofBotha and @Rabois partnered since PayPal.

English
0
0
3
1.1K
Danil
Danil@danilk·
Just launched my blog: The Power of And. danilk.substack.com/p/the-power-of… It is about building AI-native services in the real world where software meets regulation, capital, compliance, and operations. First post: why cross-functional excellence is the real AI moat. And you can read a rant version of it in the X article below. If you are building AI that touches the spreadsheet economy (legal, consulting, insurance), this is for you.
English
0
2
63
25.8K