I use node-ts. Why? I have access to every library I will need from any provider or for any task like phone number validation.

Node is fast enough, easy to develop, deploy, and scale. Anything that is slow can be ported to Go.

It is easy to find solutions to my immediate problems using LLMs.

I’m just using express and my data layer consists of neo4j and qdrant. All code is single responsibility and I use layered architecture. Easy to test and feed in to AI.

I used to work in rails, mostly to port existing code to Go or node. I found it way too messy and complicated, and our codebase was awful. I am in the minority though as I know many love rails.

Interesting as I currently work with 15 year old RoR database and a 2 years old Node project and every time I jump into the later I want to kill myself.

And I’m a mainly JS dev. Node is extremely easy to f up, so is the frontend. And with all the generated code today… it’s gonna get only worse.

> I found it way too messy and complicated, and our codebase was awful.

Are you saying that’s because RoR was used? Doesn’t seem plausible. Messy bad code is the fault of the programmer, not the language/framework

If they're a hedge fund they're probably trying to tank the US AI stocks so they can buy the dip and then in a few days/weeks it is back to business as usual.

I don't personally buy their story, and after having used Deepseek it kind of sucks and hallucinates a lot if I'm being objectively honest.

I mean a few million for this is okay - that's cool.. but it is useless. I can understand billions of dollars into something that actually works >50% of the time.

If you’re expecting to pop a bubble I think you’d buy options ahead of time to take advantage, instead of waiting for a recovery that may never come.

Well, nvidia stocks already up 10%

I know it is too early, but I'd not be surprised if this was CCP intervention using a hedge fund to try and tank US AI stocks for a specific reason.

I mean again, just being objectively honest, Deepseek kind of sucks and is maybe on par with early-2023 era models.

System: "You are a senior staff security engineer and researcher. Use the latest techniques and the internet for any new techniques to find any vulnerabilities in the code below. Do not give any explanations, only the fixed code"

User: { AI generated 1600 line handler "getUsers" with sql injection, n+1 queries, an exposed secret, plaintext passwords and credit cards, and lots of unused variables }

Congrats! Thats wonderful.

That was a good read. What impressive perseverance, and also a positive outlook. But, reading it made my FOMO worse. I guess one has to experience it fully to apply it. Or maybe not.

I quit my job a year ago to pursue my own ideas. Like the author, I had made some money from a IPO. The post-IPO environment with heavy bureaucracy killed the vibe and I quit. Note that it wasn’t a Google tier company, but I made enough money that I could take a risk for some years without income.

For me it has only been a year. But I feel like a huge loser at times. In no scenario will I give up until I’m utterly defeated or captured the castle - I’m not a quitter. But this feeling gnaws at my soul.

I have failed to find a co-founder. Most of my friends are not interested in startups. So, I have had to look elsewhere. Between people who screwed me over equity, people who didn’t want to do the work, and my own imposter syndrome (I’m from a low ranked state college, and people I meet in SF are elites), I feel horrible and alone. I tried and even successfully built some projects over the past year, but ultimately nothing worked out.

Now, I’ve decided screw it. All the advice about sell first and get a co-founder - sounds great on paper. And I have some doubts around that whole belief but that’s not important.

I know actual pain points I’ve faced in my decade long career. I’ve done both highly technical and highly functional work. I’ve talked to enough people and validated my idea. I wish this thing existed and that I’d worked on it sooner instead of trying something different with randoms.

So, I’m just going to go and build it. I like building things, and I’ve gotten enough recommendations and performance reviews that point out that I’m good at delivering value for the customer and making sure my team succeeds.

So I’m building a product. I am able to dogfood my own product as I build it which is great. It’s a somewhat complex, b2b AI enterprise product, and having built multi-billion dollar enterprise software and delivered multiple multi-million dollar implementation projects, I feel that there is really no rush. Most enterprise software and organization processes are awful and not well engineered nor well thought out.

If this fails, whatever. I tried. If it works, amazing. I like building good things and have a track record of doing so, albeit for others.

But I fear I will keep feeling like a failure until it is successful. On the other hand, it would be very cool to build a solo founded enterprise product and deliver actual success.

Anyway, thought I’d share my thoughts it feels good to type it out. Thanks to the author of the article for sharing.

The one thing I have learned on my journey through 6-7 startups is never start something without a sales person onboard. When we do we end up building something interesting with no customers. I am sure you can be your own sales person, but I don’t have it in me to do both.

Agree on the sales! Definitely need to do it.

I did hundreds of cold calls for someone else. Honestly the most intimidating thing I've done but ended up learning to enjoy it.

I also enjoyed meeting our customers at our conference at previous job and it was a fun experience talking to people and showing off what we'd been working on or understanding what we could improve.

I am not an extroverted person, so it definitely does not come naturally to me - I'm suppressing butterflies if I'm being completely honest.

My friend is on h1b. Can he join me as a co founder?

He has done zero work/contribution so far because he is tied to his sponsor.

I would like to have him as a co founder and would like any advice you can offer. Thanks!

It's possible but there are a lot of factors to consider/issues at play which are impossible to go through in this forum so you and he should consult an immigration attorney.

In 2011, we hired someone on an H1B at the startup where I worked. It was a straightforward process. The paperwork added a couple weeks to the process, and IIRC it cost $10k or so in legal fees, but for a great hire, it was definitely not a meaningful barrier.

We were already incorporated as a C corp. I would imagine that some sort of incorporation will be pretty much a requirement.

(IANAL)

Thanks. It's more complicated when the H-1B beneficiary is a founder/owner.

This was a fun read. I’m not a AI expert by any means. I’m also ESL. Please bear with me.

However the inaccuracy threshold seems fine for a museum, but in enterprise operations inaccuracy can mean lost revenue or worse lost trust and future business flow.

I’m struggling with some more advanced AI use cases in my collaborative work platform. I use AI (LLMs) for things like summarizations, communication, finding information using embedding. However, sometimes it is completely wrong.

To test this I spent a few days (doing something unrelated) building up a recipes database and then trying to query it for things like “I want to make a quick and easy drink”. I ran the data through classification and other steps to get as good data as I could. The results would still include fries or some other food result when I’m asking for drinks.

So I have to ask what the heck am I doing wrong? Again, for things like sending messages and reminders or coming up with descriptions, and finding old messages that match some input - no problem.

But if I have data that I’m augmenting with additional information (trying to attach more information that maybe missing but possible to deduce from what’s available) to try and enable richer workflows I’m always being bit in the butt. I feel like if I can figure this out I can provide way more value.

Not sure if what I said makes sense.

> Not sure if what I said makes sense.

Not sure either. But here is the lesson from this and other sources. To improve the output use multistep approach. Get the first answer, one or more, and pass it through the second verification step(s). Like 'for this * is this *' relevant? Or is it correct, does it solve the problem, etc.. Then select the answer with the best scores on the filters. You see, it's very similar to that in the original post. Get first candidates, filter.

Any way to speed it up? So of course I can try (keyword) to improve the results, but it all comes down to how well the embedding is stored, the quality of the actual user content, and of course my prompt.

I have a multi-step workflow already, but it is getting slow now going through all those steps. And sometimes, if the result is wildly incorrect, it feels really bad.

Again, I'm not an expert in this field but I am trying to learn and improve my product.

You can try using online services with best models. With prices like $1 for 10K requests it's hard to make local competitive solution. I'm going to use this approach in my personal projects.

Great! Same I’m using OpenAI models, not local. What about you?

Remember when blockchain was going to be used in everything?

Now we have AI + the ever favorite avatars.

Except AI is kinda already in everything from iphones to windows to google? Don't think that point holds well

Except has been pushed down into places where it does not make any sense just for the sake of riding the ai race. It’s like Clippy but integrated everywhere, eg in the chrome developer tools, nobody wanted it. On iPhone? I won’t be surprised if most users don’t use any generative ai feature at all. And other useful features were already inside the system but since they’re not “generative” they’re getting marketed as such or wrapped by a useless natural language prompt

I agree about the general silliness of most of the "generative" features - I was only contesting that to me, there's a clear difference in the rate of institutional belief/backing and adoption

Sure, but are LLMs? This is like saying "hashes are already used in everything from databases to network routers, so surely blockchain will succeed too!"

Not that I disagree that LLMs _will_ be used in a lot of places, many more than blockchain, but I want to point out that there is a qualitative difference between the two.

Don't forget The Metaverse.

For me the issue has been that local models seem to be pretty bad (and it’s entirely possible it’s a user error on my end) compared to calling gpt-4o.

And even gpt-4o needs constant attention to ensure it doesn’t analyze input in a haphazard way.

I made a simple AI app for myself where I ran a bunch of recipes I’ve saved over the years through batch analysis. It tagged recipes with “kosher salt” as kosher lol. I had to baby sit the prompt for this simple pet project analysis for like 4 hours until I felt confident enough. And even then I was at maybe 60%.

Now imagine for a business application. This kind of incorrect analysis would be detrimental.

Both but I like to be the one delegating and leading.