Funny you should ask I just ended up here googling "graph memory LLM"
So yea I'm very much looking into it. I want my personal agent to grow to know me over time and my life is not bunch of disparate points spread out across a vector space. Rather It's millions of nodes and edges that connects key things. Who my parents were, where I grew up, what I like to do for fun and how it ties into my personality and strengths, etc...
To have this represented in a graph which a model can then explore would allow it to make implicit connections much easier than attempting the same with embeddings.
If you get a list of all the startups that got acquired or IPO in the last 10 years, you will find it's extremely rare the technical co-founder is still around. The staying rate for CEO is like 99% while the staying rate for CTO is more like 50% (making up numbers here but this is directionally right).
With enough scale, a great CTO can be hired for the right salary. The way I answer this for myself is two-fold:
1. I've vowed I will never be the second chair guy because I don't wanna get pushed out.
2. It's important to up-skill yourself so you can contribute more value than a glorified engineering manager by driving vision, being a headhunter of superstar engineers, among others high-value skills
This is something you learn when you actually raise venture and meet peers that have raised capital. Beyond the very early stages, the CTO can easily be replaced, the CEO is the face of the business.
Hey, OP here. I view it like this: if you were to take all technical people on the market for launching a startup and plot them on a distribution to find the "valuable in early stage" subset, you would find that 35% are good enough to get a startup off the ground. They don't need to be extraordinary programmers, just good enough but with a lot of drive.*
However, for non-technical folks that are on the market, the distribution is worse. Only the top 10% ~ 15% actually have sufficient abilities to pull it off. The remaining fall into the archetype I wrote about here.
The caliber for what's in the market is skewed thus the technical person has to be a whole lot more picky than the other way around to avoid ending up with a dead weight.
[*]: This excludes specialized skills like Deep Learning or chemical bio-engineering, etc...
If you're taking a fail fast type approach, the business guys fail faster when they can't get a tech co-founder to rally around their ideas. If you can't do that, you probably can't sell the MVP either. On flipside, the tech guys will spend month and even years polishing their 'product' without ever actually launching it with actual attempts to sale it. So they fail slowly after sinking tons of time. They'll talk about it as a side-project to minimize the appearance of failure but if they put all that work in they really do want it to succeed. They just get to the point where code is good enough and can't force themselves to transition to marketing/sales/etc.
This list is a pure fallacy of statistical understanding. Skewed data and cherry-picking while also being strengthened by our subjective biases. I would appreciate the list way more if it also included founders who applied but were rejected, or for any other reason were not invested in.
For stuff like this I recommend the book, May Contain Lies by Alex Edmans, he does a fantastic job explaining how most of the time we seek data to confirm our hypotheses, instead of seeking a hypotheses that confirms our data.
> However, for non-technical folks that are on the market, the distribution is worse.
And that's where I took issue with your post -- that you are lumping all these people as "business co-founders".
That's false. To call yourself a business co-founder, you need to have the same kind of business and management chops as a technical co-founder has technical chops.
Your post seems to be complaining about co-founders who are just idea people who don't have any business skills at all. It's misleading to call them "business co-founders". You should call them "idea co-founders" or something.
If I had to re-summarize your point, I'd put it this way:
1. Success is very little on the idea, almost entirely on the execution
2. Execution has both a technical side and a business side, and can fail from a poor execution either way.
How this applies to your target audience:
You have an idea for a business? Nice -- that's almost worthless (1). You can't influence execution on the technical side, so you need to bring execution from the business side (2).
I like the numbers -- "a waitlist of 1000 people or LOI from 20 businesses" gives a better idea what kind of execution a non-technical co-founder should be capable of.
These discussions also get wrapped up in what "non-technical" means.
It can mean management skillset.
It can mean people-networking and sales skillset.
For an early-stage startup, the former doesn't bring much value (management value literally scales with employee count) while the latter absolutely does.
The point I maybe did not stress enough on is the relationship aspect is two folds:
- Have ability to gain hard-to-obtain relationships in the beginning
- Have ability to grow the pace at which you gain those over time
It the beginning it will give you the money to play the game long enough (ie: customers and angel investors). In the short term it help tremendously in deal-making, fundraising, enterprise sales, hiring superstar employee. In the long term, you can broker insane deals like OpenAI convincing Microsoft to invest $10B and bet their AI future on you or get acquired.
PS: Microsoft is now backtracking out of that situation, but Altman convincing them to get in bed in the first place is very impressive.
This is true! I argue here that the right business co-founder who drives long term financial success to the company is super rare and super valuable. The problem is the average one you'll run into at a networking event cannot deliver on that.
They are much more "I give the ideas and I want to keep 90% equity and you shut up and listen to me because I'm the CEO" guys.
You make a good point. The shrinking headcount is not necessarily tied to mass-firing. It's more likely tied to +10M newly trained engineers entering the job market every year, but only 50 positions being opened.
Over time, with each recessions, headcount will shrink at some companies, and will not grow back to pre prior levels. Over time, the line trends downwards
I worked on a fun LLM project last year where I attempt to make them actually good at engaging on social media.
I end up shutting down the project due to challenges with the Twitter API (thanks Elon) but I thought there would be value in codifying my learnings on the experience.
I appreciate anyone's thoughts/comments/criticism on how I built the system and how it could have been made better.
How you can make this more useful is by improving the messaging on your site. I looked at your hero section and I had no idea what your software does, who it does it for, why it does it and why should I care.
Since the hero section did not grab my attention, as a web visitor, I had 0 interest in investing more time/energy into reading the rest of the page to find the thread of relevancy. So I closed the tab.
Without the additional context you've put here on HN, I would have no idea what Pitch Forge is about.
Here's how I would rewrite it.
> Hero
Heading: Generate a pitch deck for your startup that gets you funding, fast
Subheading: Our AI is trained on the best decks to create a narrative that will have investors *begging* to meet you
> How it works?
1. It takes less than 5min
2. You get an E2E product from script, to powerpoint, etc...
3. Our narrative will 10x your positioning and speed up funding journey
> Why us?
1. First-time founders struggle to raise and it kills their businesses
2. The first point of contact is usually an intro or a deck. If deck isn't compelling enough, investors won't take meeting. No meeting = no check.
3. By improving your pitch, your will get more meetings. More meetings = more checks
etc....
This is already I meaningful step-up. Now I know exactly what your software does (helps create fundraising pitch decks), who it does this for (startup founders, probably first-timers), why it does it (no money kills businesses), why should I care (I should care if I want to raise money from investors and I want my business to not die)
I agree with this assessment. The copy is a little to generalized. Just by adding the word Startup to your headline, a reader would get a better understanding about what you do. Personally, I would drop the phrase Artificial Intelligence from the headline. I don't think it's necessary. Instead you should put the benefits you listed in its place. "Create the Perfect Startup Pitch in Minutes - Fast and Stress Free"
So yea I'm very much looking into it. I want my personal agent to grow to know me over time and my life is not bunch of disparate points spread out across a vector space. Rather It's millions of nodes and edges that connects key things. Who my parents were, where I grew up, what I like to do for fun and how it ties into my personality and strengths, etc...
To have this represented in a graph which a model can then explore would allow it to make implicit connections much easier than attempting the same with embeddings.
reply