Game-changer alert! I just came across a new phone provider called Popcorn and I'm sold! They're offering a global phone plan with unlimited talk, text, and data in 150+ countries for just $69/mo with NO roaming fees! I'm talking one plan, one price, no hassle. I'm definitely considering switching my phone provider. Anyone else intrigued?
Still work in progress, and I'm a noob, so will take a little while.
I'm scraping the posts every 30 minutes, and categorizing according to whether it is just a link to another page (a), or if there is actually content(b).
a => open link, and scrape said content
b => is scraped from the beginning, open and scrape links if possible.
This effectively gives me an "enriched" database, so each week, I can use the "extra" data to do a "semantic search" like: Submissions that talk about Beauty, Spain, and Beauty in Spain, or different combinations of the topics.
(RAG https://help.openai.com/en/articles/8868588-retrieval-augmen...).
The problem with only doing this once per week, is that content, that is within a niche topic, that didn't get sufficient upvotes, gets lost. But, I want to add the "weight" of upvotes and comments.
I have my own personal recommender YOShInOn which is an RSS reader that shows me about 5% of what it ingested. If you look me up in the profile I could show you a demo.
My answer to the diversity problem is this: out of maybe 2000-10,000 items I have the system make N=20 clusters with
and instead of picking out the 300 items with the highest score I pick the top 15 items in each cluster. Everything I post to HN was selected by YOShInOn once and by me twice and I think you can see the clusters at work if you note I post articles about programming, sports, environmental issues, advanced manufacturing, omics, energy technology, etc. If I pulled the top 300 it would all be arXiv papers about recommender systems with a few "circular economy" manufacturing topics.
If you found, say, 200 HN articles on a certain topic last week you might need a smaller cluster size, maybe N=5. There are other approaches to the diversity problem in the literature too but this one is easy.
I get amazing results with SBERT embeddings on HN titles and similar short texts. There is the trouble of ambiguous titles which nobody could classify but if a title is clear enough for you to get the gist of it, SBERT probably does well on it. If you are crawling the stories you are increasing your data 1000x but you are NOT going to get 1000x better results. Here is how I do on thumbs up/thumbs down classification with just titles and an obsolete algo:
Paul, this is quite amazing.
Can I email you later?
I have been "sniffing" about this idea for a little while and I am finishing up a smaller project, this is the next thing I want to tackle.
I'll never forget the day I joined a pre-seed startup, full of optimism and ready to build something from the ground up. However, it quickly became apparent that the non-technical CEO and COO had unrealistic expectations, and the CTO's limited software knowledge left me to handle everything tech-related. As time went on, I faced a series of red flags, including the CEO's dishonesty about revenue, micromanagement, and demands to "dumb down" my explanations. The work environment became increasingly toxic, with the CEO and CTO living in the office and creating a blurring of personal and professional boundaries. I was constantly asked to create new products only to have them scrapped, and was even asked to fake data to impress investors. Despite my concerns and requests for salary adjustments, the CEO prioritized his own interests and created a culture of fear and control. After a series of demoralizing experiences, I realized that no amount of funding or potential could overcome the lack of integrity and competence in leadership, and I knew I had to leave to find a workplace that valued transparency, respected expertise, and delivered on its promises.
This is a YC Startup writing incognito. If you have a horror story, you can send it to the newsletter.
“I found my cofounder on LinkedIn. We got along well, but early on, I noticed how slow he was. He insisted on being a "full cofounder" and CTO, and I gave him the space to grow, even trusting his decision to build in PHP from scratch. We raised $1M, but progress was slooooow, and soon, runway was running out.
The worst part? He deleted all our code and data, setting us back to square one. He also never took our word for what users wanted. He ended up building what he thought was right, ignoring feedback. Eventually, we had to cut the engineering team, and the CTO took time off. But things took a turn for the worse—he refused to leave, stirred up drama, kicked collaborators off Slack, and even recorded meetings without permission.
Legal battles, deleted accounts, and a rogue CTO who just wouldn't quit—this story has it all. If you've ever faced cofounder drama, you’ll relate.”
A friend of mine raised $2.1M from a US fund in the construction space. I had also tried to raise from this fund but wasn’t successful. Later, another contact in the same industry shared his experience with this fund. They spent two months "reviewing" his data room, holding meetings, asking questions, and seeing demos. In the end, they passed and funded his competitor instead. While frustrating, that's part of the game.
Now, back to my friend. The fund led his round, and he’s currently raising his next round. They initially committed to an additional $2M. He successfully secured $4-5M from other investors and returned to the fund, asking for the term sheet to finalize the round. However, the fund suddenly claimed they had lost confidence in him and his startup—even though he had grown revenue 5x in 9 months to over $150K MRR. They asked him to reach out to the new investors to help buy them out.
This is a $250M fund (Fund III), so they’re certainly not running out of money. Needless to say, he felt completely blindsided and betrayed.
Im not sure I understand your comment?
I havent down voted this...
And a good question: "What do you do?" I guess I do accounting, recruiting, legal, product, design.
And to your advice: "Maybe learn a skill and have a functional product before selling". Learning a skill, sure, I can learn a lot of skills, python to start with, but how long time before I am as a good as a CTO should be? And "have a functional product before selling" I fundamentally disagree with that point. Rather launch mvp after mvp, iteration after iteration.
Great feedback, thanks, good karma, good energy, good luck!
