
CEO and Co-Founder @TalvyHQ | Helping Companies get AI-Enabled prev. MIT, Stanford, Link Ventures principal investor | 115K+ subscribers
Joined August 2017
- Tweets4K
- Following551
- Followers1.5K
- Likes4.2K
Pinned Tweet
new vlog up: day in the life of a startup founder.
a full day of building @TalvyHQ, start to finish.
9am to 11:17pm. six hours of meetings, features to ship for multiple customers, all while managing the rest of the team.
as a founder, you never really switch off.
even at our Wednesday team lunch, with people i genuinely love, there's a pressure in the back of my head to "get back to it." , and oftentimes those thoughts keep me up until 2 AM.
but regardless of how tired i may be, I still don't miss those laker games with my dad.
people often wonder if the grind's "worth it."
to me, it is.
but only if you remember to make the time for the people you're doing all the work for.
give it a look if you're curious.
Claude and ChatGPT each have four main models right now. here’s a basic guide on how i’d use them:
Fable / Astra: big projects you can hand off. building an app from scratch or anything that might take hours to finish.
- think: open-endedness + autonomy.
Opus / Sol: most of your actual work. coding, researching, working on spreadsheets. if you’re not sure which to pick, start here.
- think: complex problem-solving + collaboration.
Sonnet / Terra: everyday stuff. rewriting emails, explaining screenshots, summarizing docs.
- think: speed + everyday tasks.
Haiku / Luna: simple, repetitive tasks. sorting 500 support tickets by topic, pulling dates from a stack of invoices, etc.
- think: volume + repetition.
also worth knowing that free plans mostly just give you access to the lighter models.
I say this because there are a lot of people who have only used free AI and are judging the technology off of its least capable models.
try giving Fable/Astra, or Opus/Sol a genuinely hard task before deciding what AI can and can’t do.
1
2
222
I think it’s understated how much AI might serve as a mass exposé on the level of ambition some people truly have, bc it’s going to remove so many “good” excuses that people have used in the past to justify why they haven’t achieved certain goals.
A $20 AI subscription can get you the informational/executional value that would have cost thousands of dollars no more than 5 years ago.
And as these barriers continue to disappear, we’re going to find out just how many people were actually being held back by their circumstances versus how many were comfortable having those circumstances as an explanation for why they never tried.
I’m someone who has a tendency to say things as kindly as I can, but unfortunately, this is the only way I can put it:
There’s a BIG difference between wanting to accomplish something and wanting to be the kind of person who accomplishes it.
For a long time, those two things were pretty easy to conflate.
You could tell yourself you’d start that business if you had the money, build that app if you knew how to code, or pursue that idea if you had the time.
And maybe those were genuinely the things standing in your way.
But as AI makes more of those things accessible, the question increasingly becomes whether you’re actually willing to do the work.
A lot of people are going to find out just how capable they can be, but I also think a lot of people are going to learn some pretty uncomfortable truths about themselves.
I think it’s fair to say that AI may just accelerate the trajectory of where you were heading anyways.
Because even though AI might be the bridge between where you are today and where you could be without limitations, it’s very possible that a lot of people may simply choose not to cross it.
6
256
spent this sunday watching sports and building video games with different AI models, truly an elite experience.
1
87
i almost want people to push back on my decisions now.
i'll say "hey, i'm about to do this, any objections?" and i'm hoping somebody objects.
i never want to find out later that someone thought it was wrong and i never considered it.
1
4
199
A lot of AI companies are doing Elizabeth Holmes type activities in the sense that they're selling features that don't exist yet during enterprise deals:
"yes that feature is ON our roadmap"
and it's sort of a moral gray area with AI, because with the speed you can ship, it might actually exist by the time the contract is signed.
makes you wonder at what point does the ability to build something become indistinguishable from actually having built it?
is it lying if you know you can build it in a weekend?
2
6
188
at what point in time do you think we'll start to feel token inflation in the same way we've experienced gas inflation?
2
1
105
i’ve told this to my wife before, and I know it’s pretty morbid:
“as long as I don’t die, we’ll be okay. We’ll be MORE than okay.”
and I say that with both faith and conviction.
Most of what can go wrong when you’re building a company is stuff you can come back from, whether it’s a deal falling through, a launch that doesn’t land, or even the company itself not making it.
so, in my mind, dying is the only outcome I couldn’t come back from.
and the reason I can make that promise to her is because I know I will never, ever stop trying.
Whatever happens in life, I know I’ll always find a way to create value, provide for my family, and get back on my feet.
and I think that having this mindset is the ultimate competitive advantage as an entrepreneur.
Knowing deep down that failure isn’t final totally removes the fear of it, and this allows you to take risks that others are simply too afraid to even consider.
So if you’re building something right now, I encourage you to do two things:
1. Remember that as long as you’re alive, you’re still in the game.
2. Bet on yourself with the confidence that you will always find a way.
12
354
since any person can build any product with AI, trust has become the only thing that's left to sell.
1
2
86
i don't know when anything I'm doing will fully pay off.
i feel like i'm just taking daily leaps of faith that the decisions I make each day are setting me up with the best chance later on.
and sometimes the leap of faith isn't this big swing, sometimes it's just putting in the daily work, treating the people around you great, not being transactional about anything.
as an entrepreneur, you might not see the benefits of your decisions right away. but you should always know that the compound effect of doing the right thing is real.
there's this important part where you're building trust in yourself that you can do hard things, and that trust is what opens up other paths.
so if nothing has come back from what you're doing yet, I encourage you not to think about that time as wasted.
you're building a foundation right now.
And early on, that's so much more important than having the tallest tower.
i know that this is almost certainly a *bit* of a marketing play for their stock value, but tbh I feel like this is an extremely viable future that AI can help achieve and we need more people spreading this level of technological optimism in a society saturated by AI doomers.
SITUATION DETECTED: Anthropic CEO Dario Amodei says AI breakthroughs will turn cancer, Alzheimer's, heart disease, and mental illness into "relics of the past" in less than a decade, ushering in "an age of miracles."
2
71
this is why we focus so much on end-user enablement and implementation rather than just "strategy"
1
80
I've been asked before if our goal at Talvy is to automate people away.
We've had people get genuinely emotional over us showing up.
Because they think we're there to replace them with AI.
And that initial fear is something we have to address head-on every single time we walk into an engagement.
Ironically, implementing AI has been one of the most human-centric experiences I've had in my career.
When you're just building software, all you're worried about is the code, but here you're dealing with people's livelihoods.
For some, it's their sense of purpose.
So when we show up, we have to be incredibly conscious of how our presence might make people feel scared about their future with the company.
It makes building trust our absolute first priority before we put a tool in front of anyone.
Without it, the technology will never be adopted, no matter how helpful it can be.
So at first, we spend a lot more time listening than building.
Because when someone is worried about being replaced, questions about their “workflows” can feel like a threat rather than an opportunity.
So before we ask about their jobs, we ask about who they are.
We exchange names, learn their stories, and establish genuine connections.
Only then do we start asking what they enjoy about their work and what they wish they could spend less time doing.
And as we start removing those frustrations, something really rewarding happens.
The same people who were initially hesitant start coming to us with their own ideas.
They go from worrying about what AI might take away to imagining what it could make possible.
And that’s ultimately what I want our work at Talvy to accomplish.
I want people to leave our engagements feeling more confident about their futures than they did when we first walked through the door.
1
4
162
AI is creating an interesting divide between people who see it as a way to save time and people who see it as a way to expand what’s possible.
Some people are using it to make their current lives more efficient.
Others are using it to explore what their lives could look like if they weren’t so constrained by what they knew how to do.
I find the latter particularly exciting.
1
66
pretty sure that if AI totally stopped developing today, we'd still spend the next decade trying to understand everything we can get out of these models.
makes me wonder if humans will ever get the most out of technology that is constantly exponentiating in capability.
2
86
random thought but i genuinely feel like i can beat 99% of startup founders in 1v1 basketball.
someone should organize a tournament where the winner gets a $100k angel check.
3
6
225
how many people have you met IRL that actually use their ai subscription more like a close friend / confidant than a productivity tool?
i keep seeing posts about people forming deep emotional bonds with their chatbots on tiktok and it's hard to believe it isn't bait.
1
115




