PretzelJudge

By default you can’t access the latest OpenAI models unless you request access. We requested access for a very straightforward use case and never got it. We switched to Anthropic and Bedrock for that reason.

Helpful link. Thank you.

I think that when people are worried about ZDR, what they really worry about is data governance. From what I’ve seen there’s a general distrust of OpenAI. AWS may keep your data around (without formal ZDR) but the concern of governance (using your data to train without your consent) seems like it would be much lower, because any breach of contract at AWS would have potential to destroy trust in what’s already a massively profitable company, so the incentives just aren’t there.

I’m not claiming OpenAI is training on API data. Just that they don’t have as strong of an incentive not to as AWS.

> Can I ask what the product was?

Software for teachers. Teachers tend to talk to each other, which is great for distribution. On the other hand, they have very little spending power. But back on the other hand again, they don't get as many cold emails as execs do.

> Research we’ve seen shows execs still spend 6+ hours per week on scheduling.

Be very careful not to delude yourself here. No one is spending 6+ hours a week on scheduling who is willing to use your tool. That's almost a full workday. If you are at this point, you probably have a human assistant who handles this.

> Would you be bothered if you were on our 23/mo plan and still had a banner there?

Probably not. But if i were, i'd probably just upgrade to a higher plan.

I like this quote:

> First time founders are obsessed with product. Second time founders are obsessed with distribution.

I have personally grown an app from zero to 100k+ users, with no marketing budget. I don't have a magic formula, but i can say in hindsight that the first 5 users were harder to get than the next 999,995.

Here was my secret: I begged. Finding 5 people to actually use my product was soooo freakin hard. Once I got to ~100, organic growth started to take over, but things like responding to emails, fixing bugs and looking into new features right away make early users personally like you and more likely to tell their friends.

The hard part is finding the right people at the right time. I googled "Persona email scheduling" and found https://usepersona.app/, so I'll assume that is you. I don't know much about your product, but i assume you are targeting people who need to schedule lots of meetings. I have a few notes:

1) Your pricing is crazy. $660/year is a lot for automated meeting scheduling. Your comp should be Calendly which is closer to $10/month.

2) What problem are you trying to solve? Calendly seems pretty close to the perfect solution for one-on-one meetings. This seems to be for scheduling larger groups. Make that clear at the top of the page. If you are directly competing with Calendly, then just make that clear.

3) I'm not the audience, but think about who really has this problem. For me, 90% of my group meetings are scheduled within my company and we use Google calendar to find times that work. I dont do this enough that I'd feel compelled to try out a new product. Sales people schedule lots of 1-on-1 calls, but not sure that group calls is a big issue. So think about who this is really for. In general, just about the only time i will try a new product is if it provokes my curiosity (not gonna happen for a meeting scheduling bot) or it is solving a problem I need solved right now. Some products are really great, but finding the right person at the right time is a huge challenge. This is why I brought up the quote at the beginning. The right product with no distribution is very hard to get off the ground. I'd say that your early goal should be to find power users if possible... if you try to spread too wide of a net, you will probably fail. Better to have a few people who love your app than a larger group of people who think it's nice.

4) Let's say you do get a few users. Well, you are an email scheduling tool, so distribution is actually built on. On your free or basic plans, you can have a banner at the bottom of outgoing emails that says "This email was drafted by Persona, the tool for automated emails". Now you have free marketing on every email that goes out. Same way that my calendly link is free marketing for calendly. If people want that banner removed, they can upgrade their plan.

Again, take this advice for what it's worth, which is very little, since 1) I dont know your market well, and 2) what worked for me may not work for you.

There are 190+ other countries that can also approve it (or that aren’t organized enough to care about drug approvals). If there was an amazing treatment working in Lichtenstein, we’d hear about it.

You can easily simulate an early stage startup by getting together with some friends and each of you calls random people from the phonebook, tries to sell them a product that they don’t want right now and then gets hung up on. The winner is the one who persists the longest after repeated failure.

The sentiment here is right, but redis does make a difference at scale. I built a web app this year on AWS lambda that had up to 1000/requests/second and at that scale, you can have trouble with Postgres, but redis handles it like it’s nothing.

I think that redis is a reasonable exception to the rule of ”don’t complicate things” because it’s so simple. Even if you have never used it before, it takes a few minutes to setup and it’s very easy to reason about, unlike mongodb or Kafka or k8s.

No we are not. The US has a much higher GDP per capita than most of Europe, but not Norway.

Sure, if you just want someone to talk to. But most clinical psychologists are dealing with more severe cases. Would you want your suicidal teenager talking to a licensed therapist trained to deal with their case, or to someone overseas whose qualification is that they speak English?

A PhD in psychology takes 6 years to complete plus another year for licensing. A PsyD is a little bit less, but typically you pay high tuition during that time, whereas most PhD programs are free (with a small stipend). So, you make little to no money (or pay money) for a long period of time and then you are presented with a choice to take or not take insurance. If you choose the former, you:

- Get paid half as much

- Have to deal with filing claims, which ultimately becomes an additional expense, since chances are you have to pay someone to do this for you.

- Get your money later instead of now.

- Have to keep meticulous notes in case you ever get audited by the insurance companies, who can refuse to issue payments if your notes don't meet their standards.

- Have to lose patients when their coverage runs out

Meanwhile, there's overwhelming demand for therapists in many cities and plenty of people who will pay cash. I truly believe that many therapists are not in it for the money... but if they are going to make less money, let's at least figure out a way to handle the bureaucracy so that their jobs are more enjoyable.

The amount of equity isn't as important as the vesting schedule. Don't vest anything to your cofounder right away. You should vest on a four year schedule with a one year cliff. There are good reasons this is the standard.

As far as the amount, make sure the other person thinks it's fair. If they don't find the split fair, then it'll just create issues later. Better to give away 49% of your company and have it succeed than to continue to own 90% and have it fail.

It isn't you. I have a solid resume (I know because I've interviewed hundreds of people in my career), but the market is just much slower. In 2020-2022 I was getting 5-10 recruiters reaching out every week. If I thought something was interesting, I'd respond saying that I'm not really looking for anything new, but I'm open to a conversation. People would still be thrilled to speak to me, even though I was upfront that I'd only leave for something perfect.

Lately, I get maybe one person reaching out every month. I responded to one last week, even though I'm not looking, just to see what the job market is like. I was ghosted.

I think we're just not in a hiring frenzy right now. People don't want to leave their jobs and risk it on something new. Some companies are laying off. For the companies that are hiring, they are probably seeing many more applicants and a higher conversion rate.

Things will come around.

I'd bet that YC would love to fund fewer dev tools, but there's a bias towards who applies to YC in the first place. It's not that YC isn't well known outside of tech circles, but they are especially well known in tech circles, and that affects who applies.

> While it may seem like YC invests in a diversity of companies across <healthcare, banking, travel>, it is an “Airbnb for X” type of diversity... if you are not building a product that is fundamentally like an Airbnb (or alternatively, a dev tools startup), then YC will not be interested in you.

I'm looking at the S23 batch and among them, I see a product for automating processes in healthcare, an AI voice assistant, a company tracking methane emissions with satellites, a company fixing GPS errors, building decarbonization, an AI for medical coding and others. None of these feel like AirBnB to me and certainly are not dev tools.

>As a result, most YC companies look the same. While I won’t name specific companies in any of the batches, YC companies will generally... have low capital and infrastructure costs, with low marginal variable costs

Yeah, that's the kind of company that wants $500k to get started. If you are building something capital intensive, then 500k probably doesn't move the needle for you. Either you are going to start small and build slowly (which makes for a bad VC investment) or you are going to need a lot more than 500k.

>For instance, if you are pitching a research lab without a specific plan to profitability (e.g. Mistral AI), then YC is not for you.

Yeah, Mistral AI raised $113M. If you can raise that as a seed round, then you don't really need YC. YC just isn't in the business of financing things that need 9 figures to get started.

Neh... web programming is not going anywhere. Sure, the current environment is exposing all of the companies that never really had a business model other than raising more and more money. But there's still plenty of space for software to grow within existing profitable companies and new companies that can be more capital efficient.

He claims you can strip out some of the inefficiency which is related to interoperating software... that's not going anywhere. There will alwyas be walled gardens or code that breaks. Maybe that'll even increase. And even though what AI can already do is incredible, I think it'll take a long time before it can build and maintain large scale projects. In the meantime, it'll certainly make developers more efficient, but so have countless other tools.

That said, I think that remote workers are going to have a reckoning. If i'm hiring for people to be in the office, then i have a much smaller pool, replacing people is tougher, etc. But once I'm opening the doors to remote work, why not do so in a country with a much lower cost of living? I see people on HN constantly talking about how to succeed remotely, you need to be able to work in an async environment, etc. Well guess who loves the idea of an async environment? Upper managers who see ballooning payrolls and realize that an async/remote workplace doesn't need to consist of people in high cost countries.