1:04:40Greg Isenberg · Interview
Stop using Claude. Start using Codex?
A 64-minute masterclass where a Codex true believer converts a Claude Code skeptic, live, on camera.
April 27thThe argument in one line.
A solo builder with a well-orchestrated enrichment and clustering stack can produce a Techmeme-quality, personally-tuned AI news feed in a few weekends for roughly $300 in API credits, and the harder problem remains deciding what not to build.
Read if. Skip if.
READ IF YOU ARE…
- A solo builder or technical founder who wants to build a personal intelligence tool without hiring a team.
- A developer who has used LLMs for individual tasks but wants to understand how to chain them into a durable production pipeline with retries and observability.
- Someone curious how vector embeddings differ from keyword search in practice, with a concrete news-clustering example.
- A builder evaluating trigger.dev as an alternative to Supabase cron jobs or custom queue infrastructure for an AI-heavy workflow.
- Anyone interested in how to layer an editorial scoring system on top of raw clustering to get a prioritized, signal-rich feed.
SKIP IF…
- You are looking for a finished product to use — Nylon is a personal project and not publicly available.
- You need a step-by-step tutorial with code; this is a live demo conversation, not documentation.
The full version, fast.
Kevin Rose built Nylon — a personal Techmeme clone tuned for AI news — using Claude Code, TypeScript, and roughly $300 in API credits. The architecture layers RSS ingestion from 63 sources through a multi-source enrichment waterfall (iFramely, Firecrawl, Gemini as fallback), uses a judge to pick the best content per field, generates vector-rich TLDRs for embeddings stored in Postgres via pgvector, and clusters stories using a vector similarity algorithm. When three or more RSS stories cluster, the system expands coverage by hitting Brave and Tavily search APIs. On top sits a gravity engine that scores clusters by industry impact, novelty, viral potential, and PR-fluff risk — turning a raw feed into a prioritized editorial ranking. The meta-point: in an AI-abundant world, the most valuable skill is knowing what to cut, not what to build.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Who's talking.
00:00guestKevin Rose
00:00hostGreg Isenberg
Where the time goes.
00:00 – 03:09
01 · Intro and what Kevin plans to demo
Greg sets up the premise; Kevin teases Nylon and frames the challenge: AI makes building trivial, but clarity about what not to build is the real skill.
03:10 – 06:44
02 · Techmeme tour: how signal gets ranked
Kevin shows Techmeme, explains how social signal volume creates ranking, and explains why he wanted an AI-velocity-tuned alternative.
06:44 – 11:23
03 · RSS sources and the article pipeline
Walks through 63 sources in the Nylon admin, shows articles ingesting in real time, pipeline status tracking in Postgres.
11:23 – 13:01
04 · Winner selection: iFramely vs Firecrawl vs Gemini vs RSS
A judge evaluates each enrichment source per article field and picks the winner. Gemini serves as grounded last-resort.
13:01 – 16:37
05 · Why iFramely and Firecrawl
iFramely for rich card metadata; Firecrawl for deeper crawling with stealth mode. Both explained via live demo.
16:37 – 19:49
06 · TLDRs, vector embeddings, and why they beat keyword search
Vectors as mathematical meaning representations stored in Postgres. The Apple-sues-Google example. GPT-4o mini for TLDR generation.
19:49 – 24:58
07 · Task orchestration with trigger.dev
Why Supabase cron was insufficient; trigger.dev gives durable TypeScript tasks, auto-retries, execution traces, and local-to-production parity.
24:58 – 27:07
08 · Clusters: expansion via search APIs
Once 3+ RSS articles cluster, Brave and Tavily search APIs pull in additional coverage outside the RSS set.
27:07 – 31:31
09 · The gravity engine: editorial scoring rubric
Multidimensional scoring: industry impact, novelty, viral potential, builder relevance, PR-fluff risk, visualized as a 2D matrix.
31:31 – 34:53
10 · Product management: gut, iteration, cutting
Kevin's build philosophy: one feature at a time, use AI as a sparring partner for architectural choices, discard failed branches fast.
34:53 – 37:03
11 · Synthetic audiences and personal software
Toby's Shopify synthetic audience idea; Kevin's counter — build for yourself first, then find the 500 people who want the same thing.
37:03 – 43:52
12 · What success looks like
Maybe never launches publicly. Real goal: a personalized signal surface filtering by interests and trusted-person influence weight.
43:52 – 47:19
13 · Retention mechanics and IdeaBrowser
Greg demos IdeaBrowser's daily email mechanic. Kevin: retention comes from genuine relevance, not tricks.
47:19 – 50:34
14 · Blurred presence blog project
Kevin shows a second side project: a blog with a blurred webcam silhouette providing ambient human presence. Built in Claude Code from a 12-year-old idea.
50:34 – 51:55
15 · The best time to build
Kevin on aphantasia and how AI filled the gap. Vibe coding criticism dismissed: if 50K users crash your server, that is a great problem.
51:55 – 56:25
16 · DIG, True Ventures, how to connect
Kevin's Venice studio open-door policy; VC philosophy — talk everyone out of raising capital unless they truly need it.
Lines worth screenshotting.
- The most valuable builder skill in an AI-abundant world is knowing what to cut, not what to ship.
- A solo builder can replicate Techmeme-level curation for roughly $300 in AI credits over a few weekends.
- Vector embeddings solve for meaning-level grouping that keyword search cannot — Apple sues Google and Google sues Apple share the same keywords but are opposite stories.
- Using a judge to pick the best enrichment source per field (iFramely vs Firecrawl vs Gemini vs raw RSS) is cleaner than trusting any single source.
- trigger.dev gives solo builders the same durability story — retries, traces, local-to-production parity — that enterprise teams build from scratch with queues and workers.
- When 3+ RSS articles cluster on the same topic, hitting search APIs for expansion turns a finite RSS corpus into a near-complete picture of coverage.
- Detecting paid-sponsorship PR clusters via vector distance is possible: if 20 articles share the same key points and publish within the same hour, the inter-article distance is suspiciously tight.
- Building for yourself first then finding who wants the same thing has produced some of the largest consumer internet products — Digg, Zero fasting app.
- 500 genuinely enthusiastic users is a more meaningful signal than millions of passive ones — the internet has warped our scale intuitions.
- Novelty scoring over a long time horizon is the signal most worth building for: the first Hacker News mention of Bitcoin looked silly but the novelty score would have been off the charts.
- Vibe coding criticism misses the point: if the app crashes under 50K users, that is a distribution problem, not a code problem — and that is the good problem to have.
- The era of VC as MBA gatekeepers who lecture founders is ending; the next generation of useful investors are themselves building in public alongside founders.
Build the pipeline, then cut until it sings.
WHAT TO LEARN
The real lesson is that modern AI tooling makes sophisticated pipelines tractable for one person — and that tractability makes discipline about what to cut the decisive variable.
01Intro and what Kevin plans to demo
- The question worth asking before starting any AI project is not whether you can build it — you can — but whether you are clear enough on what not to build that the result will be usable.
02Techmeme tour
- News aggregators derive their ranking signal from volume and social weight, not editorial opinion — a lesson applicable to any feed-based product design.
03RSS sources and the article pipeline
- RSS is still alive and ingestion is cheap; the complexity lies in enrichment and quality normalization across dozens of sources with inconsistent data quality.
04Winner selection
- Trusting a single enrichment source produces inconsistent quality; a judge that evaluates multiple candidates and picks a winner per field gives the database the best available representation of each article.
05Why iFramely and Firecrawl
- Two distinct crawling layers solve different failure modes: metadata APIs for structured card data, and deeper crawlers with stealth modes for sites that aggressively block bots.
06TLDRs and vector embeddings
- Vector embeddings capture semantic meaning that keyword search cannot — the same words in a different order can describe opposite situations, and only embedding-based retrieval catches that distinction.
- Generating a purposely verbose TLDR solely for embedding quality is a useful pattern: the richness of the input text directly improves the quality of the resulting vector.
07Task orchestration with trigger.dev
- Background job durability is not a luxury for solo builders; it is what allows the pipeline to keep enriching data while development continues locally.
- Lightweight background tasks cost far less than perceived: thousands of daily executions can run for under $100 per month.
08Clusters and search API expansion
- Treating a cluster threshold as a trigger to query external search APIs turns a bounded RSS corpus into a near-complete view of any given story.
09The gravity engine
- Building an editorial scoring rubric as code — with explicit dimensions like novelty, actionability, and PR-fluff risk — externalizes editorial judgment into something inspectable, debuggable, and improvable.
- Detecting undisclosed paid media is tractable with vector embeddings: coordinated campaigns produce suspiciously low inter-article distance scores.
10Product management and cutting
- Treating AI as a sparring partner for architectural decisions accelerates technical choices where there is no obvious answer.
- Throwing away a failed branch of work is a feature of a healthy build process, not a failure — the learning is the asset.
11Synthetic audiences and personal software
- Personal software — built to solve your own problem at full quality — has repeatedly produced products with larger audiences than market-researched ones, because genuine frustration produces genuine solutions.
12What success looks like
- A tool that surfaces the five things worth your attention from 1,000 signals — filtered by your interests and the influence weight of people you trust — is a different product category than a news feed, and it requires personalization infrastructure the feed itself cannot provide.
13Retention mechanics
- Email-driven return mechanics work when the email delivers genuine value — a daily idea, a curated signal — rather than a notification that the product exists.
14Blurred presence blog project
- Ideas from a decade ago that were technically impossible then are now buildable in a weekend — revisiting your old idea backlog through the lens of current tooling is a high-leverage research move.
15The best time to build
- AI tools fill in individual skill gaps — syntax retention, visual pattern-holding — that previously blocked non-traditional engineers from shipping production software.
- The primary barrier to shipping is finding something people want, not code quality; a buggy product with 50,000 users is a tractable engineering problem, while a polished product nobody uses is not.
16DIG, True Ventures, how to connect
- The most useful investors for AI-era founders are those who are also building — they can pressure-test ideas from experience, not theory.
Terms worth knowing.
- RSS
- Really Simple Syndication — a standardized feed format that lets sites publish content updates as machine-readable XML, making it easy to ingest many sources without scraping.
- iFramely
- A link metadata API that returns rich card data (title, description, images, embed codes) from any URL, similar to the preview cards shown on social platforms.
- Firecrawl
- A web crawling service with AI-assisted extraction and stealth modes that can bypass aggressive robots.txt blocks to retrieve full article bodies.
- Vector embeddings
- Mathematical representations of text stored as high-dimensional numeric arrays; similar-meaning content clusters together in vector space, enabling semantic search and grouping that keyword matching misses.
- pgvector
- A Postgres extension that lets you store and query vector embeddings directly in the database alongside regular relational data.
- trigger.dev
- An open-source background job platform for TypeScript that runs tasks durably with automatic retries, execution traces, and observability — replacing fragile cron jobs.
- Gravity engine
- A custom scoring layer that rates news clusters across dimensions like industry impact, novelty, viral potential, and PR-fluff risk to produce a single editorial priority score.
- Grounded search (Gemini)
- A Gemini feature that turns on live web search to retrieve and summarize current article content, used as a last-resort enrichment fallback when crawlers fail.
- Tavily
- A search API optimized for AI agents that returns clean, structured results suitable for programmatic consumption.
- Aphantasia
- A neurological difference where the person cannot form voluntary mental images — Kevin Rose disclosed he has it, which explains why syntax-heavy coding felt inaccessible before AI assistance.
Things they pointed at.
12:00toolGemini (Google)
27:13toolBrave Search API
19:11toolOpenAI embeddings (large model)
15:59toolGPT-4o mini
37:44toolVercel AI Gateway
06:22productDIG (social news reboot)
52:00toolClaude Code
Lines you could clip.
02:40
“The future skill is not what you build as much as what you don't build.”
18:04
“There is a huge difference between Apple sues Google and Google sues Apple — and that is impossible to do with keyword search.”
37:47
“If 500 people were standing outside your house cheering you on, you'd feel like the biggest rock star in the world — we lose perspective of what success means.”
31:25
“I actually just uncommitted that entire GitHub PR I spent four hours on and threw it away. That's all building is — failure after failure. Failure is awesome because the next time it's gonna be a little bit better.”
51:24
“I found out I have aphantasia — I can't visualize things in my mind's eye. I always thought people were joking when they said picture a sheep jumping over a fence. Now AI fills in the deficiencies wherever they are for you.”
Where the conversation goes.
00:00 – 06:44steadyProject overview and philosophy
06:44 – 16:37denseIngestion pipeline and enrichment stack
16:37 – 19:49denseVector embeddings and semantic meaning
19:49 – 24:58denseOrchestration with trigger.dev
24:58 – 31:31denseClustering and story expansion
31:31 – 47:19steadyBuilder philosophy and personal software
47:19 – 56:25sparseIdeaBrowser, side projects, wrap-up
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
00:00Today's episode is a show and tell episode with none other than Kevin Rose. He takes us through his entire AI workflow and lets us in in a new product that he's developed that he hasn't shared anywhere, and we learn about how he thinks about building new products, Claude Code, Vercel, all the tools he uses.
00:20So it's super, super fascinating. Kevin is just one of those iconic entrepreneurs and to have an inside look into his AI workflow and how he's building products in this AI age is absolutely fascinating.
00:32It got my creative juices flowing. I think it will get yours too. And if you stick around to the end of the episode, you will understand how to build products in the AI age, how to think about it, and what tools to be using.
00:46Enjoy the episode.
00:55We got the one and only Kevin Rose on the podcast. I'm super excited to have him. He has been going down the AI rabbit hole for a while now, and I wanted Kevin to come on just to share, you know, what are the tools he's using, some of the AI workflows that he has, and he's just gonna screen share.
01:17Kevin, by the end of the this episode, what do you think people are gonna get out of it? Like, if they stick around.
01:22Yeah. Well, I certainly first thanks for having me on.
01:26It's I love the content and everything you're producing. It's so important right now especially with things moving so fast. But I think at the end of it you will see that things that are technically out of bounds for you.
01:38Like things that you just think that you cannot do are very much possible as a solo engineer, solo designer. And now we're finally at the point where I'm not even calling it slop anymore.
01:51Like, call the AI slop or whatever it may be or however you wanna look at the code. It's damn good, and it's getting better by the week. And I think that you'll just I hope that you'll be inspired to go build something amazing because I'm gonna show you something that is a little mad, science y, weird, and all over the place.
02:09And you'll get to see kind of a a raw version of my brain and how deep I go on some of this stuff. Um, but I think that's the beauty of it is is is this idea that we can go build anything and oftentimes when we do, it ends up being a a little bit of a I don't know.
02:25A sandbox that can be a little too big and messy and then you have to refine it back to something that's actually usable. So I think that the future engineer and the future developer and the future product builder here, it's not gonna be what you build as much as what you don't build if that kind of makes sense.
02:43Because it's gonna be so easy to build anything and everything to pair it back to something that's really usable, think is gonna be the a real skill.
02:52Yeah. The hard part is the clarity. Like, how do you get the clarity to know what to build?
02:58At least that's that's that's what I've been struggling with. Yeah. And then that's actually rolls right into the the project that I've been kinda working on just for fun.
03:06And yeah. So I'm excited.
03:09Alright. Let's get into it. Okay.
03:11So we can I'll show you kind of this, you know, quote unquote vibe coded or let's just call it coded project. And I'll tell you the inspiration and then we'll get into some of the the kind of dirty details and how it functions and what's possible as someone that kind of walked into this thinking.
03:29Actually just asking myself, can I pull this off? So let me go ahead and do a little screen share here.
03:38Alright. You should be able to see Techmeme up there. Right?
03:41Yep. Okay. Awesome.
03:43So first thing I and and I hope we don't, uh, in the edit, this has to make it in. I have nothing but a huge amount of respect for Gabe and what he's created in Techmeme.
03:55Uh, what I'm building today is not meant to be a competitor. I don't plan on launching it in its like form. It was mainly a personal curiosity of can I build something that's on par or better than Techmeme by myself and call it like a week and see what that would look like?
04:12And one of the nice things about Techmeme is, for those that don't know or for those that are lightly familiar with it, it's been around for a long time. Gabe's been building software since I was back in 2004, which is crazy.
04:24We were both kind of you know, experimenting in early social news and what that meant. And what you see here is an aggregator that pulls from RSS and other sources.
04:33I don't know all the sources that he pulls from, but then also pulls from social media as well. And when you gather momentum, whether it be multiple news stories coming in or multiple people tweeting about something or talking about it socially, that is considered signal and then that is considered used as part of, you know, ranking and showing you what's prominent here.
04:55And so what what what you're seeing is actually visually, the higher an object is here, the more weight it carries with a user.
05:05Because if you come in here and you scroll down the bottom and you see something with very little tweets on it, it doesn't nearly carry as much impact as, you know, something with 15 or 20 different x posts underneath it. And so I like what he's doing here.
05:19I think it's a it's a really cool way to say, you know, maybe you might recognize some names, especially in tech given how small the ecosystem is. You can probably look through this list and say, oh, I know these two or three people. They're talking about it.
05:29What are they saying? And might I just hover over and see what their comment was about a particular story? So a lot to love here.
05:36I mean, Gabe's done a a great job at at creating this, but I, um, have a slightly different area of interest. A lot of this is big tech news, and I kinda wanted to dive into, like, more of what's going on in AI because AI is moving just so fast.
05:49You know, how can I slice slice this in really interesting ways to find my version of this? And this is stuff that we're playing around with that dig with the reboot as well. So it kinda goes hand in hand with some of the kind of exploration that we're doing in a lot of these different areas.
06:03And so my job these days, you know, with Alexis Ohanian on at at DIG is we have Justin as the CEO who runs kind of like the day to day and and builds out, you know, the Reddit competitor version of DIG. And then we have, you know, me that I'm kind of more in the labs area where I'm pushing on the edges of ideas and saying, what might we do?
06:23And does any of this make sense to roll back into the main product? And then Alexis is kind of just overarching, you know, great ideas, so much depth of knowledge around, you know, both micro and macro communities and how they work and how they scale and what to do and not do and some of the missing tools.
06:40And that's kind of how it all comes together. So long story short, this is what I built.
06:46So first thing I wanna show you is that there's gonna be some errors here because it's been a minute since I have actually done the whole actually touched this code base.
06:58But what you see here is I I have 63 sources of information coming in. And so these are from just your classic RSS feeds. Yes.
07:06RSS still does exist. A lot of sites do break RSS. And so there are actually ways to go and add in kind of scrapers that will create dynamic RSS for you even though RSS is no longer around.
07:20Um, but you can see here, um, it's a bunch of different, uh, RSS sources that are coming in here. Some of them are Reddit, some is TechCrunch, but all largely tech related.
07:32So because this hasn't been running a while, we've got a lot of jobs that have to run-in the background here to go and recrawl all of this.
07:44So I haven't run this project in a few days. And so it's essentially going out and hitting all these different RSS feeds, pulling in and ingesting all that information, bringing it into the engine, and then expanding upon and there's orchestrators and there's ways to actually go in and and resolve certain chunks of information and find out more about individual stories.
08:05And I'll get into that in a second, but it we're playing a little bit of catch up here. So if the news stories for some reason don't look exactly fresh fresh, you'll know why.
08:13But they they'll they'll be pretty damn close here in a couple minutes. It just needs to go and, uh, and do this this catch up. So, um, going back to the RSS feeds here.
08:25So okay, great. We've got sources. Now what do we do?
08:28Well, sources lead to articles. And when we go over to articles here, can see now it's starting to pull in some of these articles in real time. So we've got Mac rumors thirteen minutes ago and then we've got a little status check here and the status check is really what happens once the article is pulled in to the system.
08:46Like how do we process that article? How do we figure out what to do with that article? So you twenty one, twenty five minutes ago, twenty nine minutes ago, The Verge, nine to five.
08:55And so we're pulling in and saving all of this in a Postgres database. And then we're also going in and doing things like figuring out who the authors are. And this comes down to author reputation.
09:06And so we can start and look taking a look at different authors that are here. And I can click on, you know, any of these authors and actually, you know, see what articles they've actually contributed as well.
09:18So, you know, this individual art author here and what are all of the articles that they published. But what's more interesting here is actually what we're doing with these articles. So what happens is is as I start to pull on these articles, and this is where it gets pretty crazy because, you know, I didn't have this didn't really know.
09:43I understood technically how to pull this off in terms of the the tools that would be required, but I couldn't write the code to do it. And so once we go in and we take an individual article let me just get into an actual, um, an article itself here.
10:00Let's see if that will take me into the actual that might pop me out to okay. Here we go. So here's a article right that now that just came in and talks about the new AirTags that just got announced.
10:11And this came in from MacRumors. It's got an ID associated with it. So I was setting all this up within the database.
10:18It has a cluster membership associated with it, which we can get into in a little bit. But also it has a pipeline status. And so what happens is is it comes in via RSS.
10:28I hit iframely, and iframely gives me back additional metadata about it. So like title, description. Sometimes it can pull in even additional kind of deeper context about it, article data, things like that.
10:39And I want to store all of that. That's really important for me to have that rich data. And then I can see that Gemini has been running against it as well, which can get into in for in a minute for as to the reason why.
10:50And so the winner has been resolved. The TLDR has been generated, and the embedding has been done. So the winner is basically saying, okay.
10:59There are a bunch of different factors here that might come in. Like, for example, if I'm pulling this in with RSS, I might not get the description there because RSS sometimes is like kind of truncated.
11:09I might not get a really good description. But if I go and I hit iframelie, then I might actually get a longer and better description.
11:16Or if I go and hit fire crawl, I actually might get the full actual paragraph of the or the full body of the article. So what we do is we have a judge that looks at it and says, okay, which is the best of these three that came in and pick a winner? And the winner is the one that actually gets stored to the database.
11:32So that's what winner means. We do that on a on a variety of different things. We I don't know why we say we.
11:36I I do that. I'm not the one that actually ever coded this thing. Um, so basically, can see here that for this particular, uh, story, it did not have I think we got a failure.
11:49We're coming from Mac rumors. The winner was actually Gemini.
11:53So Gemini is used as a last resort. So you can see here FireCrawl returned a four zero three. We had a hard air there.
12:01IFramely, we had some kinda low quality signals. We got a successful crawl.
12:06We had low quality signals. And then RSS, good it's it's used as a fallback.
12:12Good amount of characters here. We can see what the actual summary of that content was and we got ours out of our SS. But as you can see here, it's actually pretty crappy RSS poll.
12:21There's not a whole lot of data there. You know? It just doesn't look that well or it doesn't look that good.
12:25So what we did is we hit Gemini and said, hey, Gemini, turn on your ground truth, turn on your search, go out and figure out what is like what is this actually about. Right?
12:37And oftentimes, they won't come out and say this, but they kind of crawl the article and give you back what the article actually says. And they might reword it a little bit, but it's largely what the article says.
12:48And I can prove that because they can you can do that with Reddit articles and things like that where most all everything will fail because Red's so hardcore about blocking it. Does that Greg, does that make sense so far?
12:58That does make sense. I just have one question around, like, why did you choose FireCrawl and iFramely,
13:04especially for and can you just explain what the what they do for people who aren't familiar with those services? Yeah. For sure.
13:11So
13:13let's go ahead and take the actual article here. And so I'll show this tab instead.
13:21And so you can see this is the MacRumors article. Right?
13:25Yep. Now, if I take this Mac Rumors article and I copy and paste it and then I go into iframely here, I can paste in a URL, any URL.
13:36And obviously, I'm doing this via API, but I just want to show you how it works. And I hit check URL. And what it's gonna do is it's gonna give me this this card back.
13:44So it's almost like those cards that you see on on x, you know, where you get a piece of rich media, you get the best possible image, get a beautiful title, nice little description here. Sometimes it's a little bit longer.
13:55And then I can just jump in here and look at the JSON and see like, okay, what did we get here and kind of expand out and say, do do we get a high quality logo? Like, you know, like, I can look for and parse through all of this and try and find high quality data. And then also, I mean, it's called iframely largely because if there is an embedded piece of media, you also get all of those embed codes as well.
14:16So you get all of it. So I can actually take and the reason I like Gemini is because they're tied in with YouTube obviously and I can get transcripts back from full YouTube videos, which is great for vector embeddings when I wanna figure out what it's actually about.
14:32Okay. Does that make sense? Crystal clear.
14:35Okay. And so you can consider fire crawl to be a very similar thing. It is a little bit more on the crawling side, so a little bit more fine tuned around crawling.
14:44They have some AI aspects to it as well where AI will actually try and go out and figure out how the best way to kinda scrape the content. And then, yeah.
14:54Lastly, with fire crawl, I think it's just they have some stealth modes that you can turn on. So some of these news sources, they get really picky about what they what they're allowed to be crawled. And so you can kinda like hide and turn on stealth mode and then get to the actual data.
15:08And it's not my my intention is not to like, you know, get around their ads or do something like evil here. It's really just to figure where's the signal and to and to start to cluster these things together. So back to, uh, and then by the way, the working title for this is my my little thing that I code on nights and weekends.
15:25It's called nylon. It's my little incubator that I I work on. So that's, uh, that's why you see nylon up here in the corner.
15:31Alright. So scrolling back down, we can see that okay.
15:35So here's the resolve content iFramely. We got the picture. We got the author.
15:39We got the summary. And here's the main content from Gemini. So we use that Gemini one here.
15:44IFramely won the summary. Gemini one the actual main content.
15:50And then what I hit is I hit GPT five mini largely because it's fast as hell and cheap and you don't it's very it's quite smart. Anthropic has a handful of models that also fall in this camp. I mean, they all do.
16:03You know, it's really at the end of the day because I'm using vector embeddings from, you know, OpenAI, it's just when I have one model provider, I just stick with it.
16:14And unless I really need to bounce around, which I did for the actual Gemini crawl and understanding of other things. So I try to keep it as simple as possible, but sometimes you need multiple models.
16:23And I will say Vercel's AI gateway is a great way to kind of code once and just flip a model on on the go. So, um, highly recommend checking out Vercel AI gateway as a way to quickly swap models, uh, rather than having to recode it.
16:37Alright. So I want a TLDR. That's important.
16:39I just want that human readable so you can see that. So I want a TLDR that is a vector TLDR. And so for people that don't don't know, vector embeddings are really interesting.
16:50I had never worked with them before. I'd heard about them. I understood technically how they function.
16:54But the point is that if you take a keyword rich and kind of deep understanding piece of content, you can create mathematical representations of that content and embed these with OpenAI and store them in Postgres by using a vector extension.
17:10So they're actually stored in your database as of those pieces of math. And when you apply some of these clustering algorithms on top of them, they get really good at nuanced information where keyword search would completely fall down.
17:24So the old school ways back in the day in 2004 when I launched, you know, Dig as a social news site, if you search for, you know, Apple releases whatever, it's it's fine.
17:38It's looking the word apples, looking for release, and it's looking for whatever it was back then that they were doing like a iPod or something. Right? And it would it just does it based on just can I find that text in the database?
17:49And if so, show me back the article. The beautiful thing about what we have today with our understanding of linguistics and around using vector embeds and algorithms on top of that is that you can say there is a difference even though they're both have the same type of keywords.
18:07But there is a huge difference between Apple sues Google and Google sues Apple. And that is impossible to do with keyword search because you're not understanding at a deep level what's going on here.
18:22Right? So anyway, this is a very rich, purposely rich, uh, longer form version of the TLDR used just for for vector embeddings.
18:34And then I also wanted to create some key points here that we can use to feed into other models later when we're comparing the difference between articles when we see multiple articles starting to get clustered together. And then I don't use this, but I asked AI like, hey, write me a like a spicier title, like a title that people might click on more or find more interesting and just to kind of rewrite the title.
18:54And and Techmeme does this too. Like, when you go to the front page of Techmeme, it's not the title of the article. It's like actually what their editors chose to write.
19:00So I just wanted to see how this looked. And then I wanted to put it in one of, uh, three different categories, um, tech core and a couple other categories largely because there's a lot of stuff that come through these tech feeds, especially when you add in like Forbes and some of these others where it's like, you know, does not relate to actual core tech or AI or the things I care about.
19:17I just want to put them in a bin. And so here's the embedding. I'm using the large model from OpenAI.
19:24You can see when it was generated, when it was done. And then just some information about making sure that we don't recall the article and have all that. So that's done.
19:34Okay. Any questions so far? We got one article into the system.
19:39Uh, no. Keep going. Keep going.
19:41So
19:42now let's talk about clusters. So clusters, uh, well, actually, let me show show off one other thing.
19:49So how does this actually work? So how does that all of that work on the back end? Right?
19:52And there's a couple different ways that you could do this. You could say, I wanna kind of, um, write this in Next.
20:00Js, and I wanna have this just be a function. And if you wanna get a little bit more fancy, you could say, okay, I'm using Supabase, so I'm gonna throw on, uh, some cron jobs there and fire off these things.
20:12And I don't know if they fail or, you know, there's there's a lot of, um, it gets a little dicey here because oftentimes things bad things can happen.
20:22Like, you can have a RSS feed that gets blocked temporarily. You could have timeouts. You could have a model that actually doesn't complete.
20:29Um, and so I need durability around a lot of this stuff. Right?
20:33And so what I do for that actually, let me just do a screen share again, is for the durability side. Okay.
20:40I use a service called, uh, trigger dot dev. And so I like trigger dot dev because what it does is it allows me to create, um, these functions, and they're all TypeScript that live in the cloud and that they are fired off either when I call them from my app or at a certain cadence.
21:02And so there are these orchestrators that will go in here. You can see the expansion orchestrator.
21:08That's when it go wants to go in and expand a story and actually see, you know, more more information about it. You can see clustering different things here.
21:16Here's my fire crawlers, my Gemini, here's my Framley. So anytime I fire something off to be enriched, I actually create this new little micro instance that goes out and runs on its own.
21:29You can see here the ones that are executing. See the little spinners here? So these are all kind of running in real time.
21:34And then you can see the compute charge to go off and execute these and how many milliseconds they they took. Now the thing that's interesting here is you actually get to see the whole chain in which everything went down here.
21:48So you can see, okay, I was looking to kind of resolve a story ID, locked in candidates. I got, you know, one picture, one summary.
21:57Um, did it resolve the winners? Yes. Gemini was the winner here.
22:00Uh, it was published at and then it it finished the whole process. The nice thing about this is that if something fails, I get retries for free.
22:12And so I will automatically you know, if an AI TLDR for a vector embedding, which is very important as we're building clusters, fails once, fails twice, it will continue to retry as, you know, and automatically spin up these instances and try again.
22:28And then will report back to me via Sentry or any other type of monitoring software that I have on the back end that, oh, I had a failure. Why did this actually happen? Right?
22:37And so I like that because a couple things. If I'm developing this locally and I don't have this on production, my data continues to be enriched in the database.
22:46And so I can continue to develop before I actually deploy deploy to production. So when I'm running this locally like I'm showing you here, it's like it's it's it's functioning and continuing to build things out. There's a few things that are are still tied locally that I need to catch up on to fire those back up because I just don't wanna be burning through cache for for no reason if I'm not using it.
23:06But yeah. So does that make sense?
23:09Yeah. I also think
23:10if I remember correctly, trigger dot dev is open source, so we'd like that. I also think it's relatively cheap.
23:19Oh, yeah. It's really inexpensive. Like, it's something like $10.15 bucks a month for 50 tasks or something last I checked?
23:27It's well, I mean, I I think we look when we looked on there, we were actually seeing the per per I mean, I'm running thousands and thousands of these. And, know, I yeah.
23:36It's like under a $100 a month and that's I mean, when I say thousand, I mean like per day. Yeah. So it's it's it's not obviously, if they're gonna be super you you can choose your instance type like with anything else.
23:47These are really lightweight non computational kind of tasks to do. Some people use trigger for things like, you know, using FFmpeg to do encoding and and, you know, things of that nature that are gonna require a larger instance are obviously gonna be more more pricey.
24:04So it really depends on like everything in in cloud on on the workload. Right?
24:09Totally. Yeah. Just don't want people to see this and be like, my god, this is probably thousands of dollars a month.
24:13It's like shockingly inexpensive for what it is. Yeah. And the other thing I was gonna tell you is that
24:20Vercel now has something called workflows that they've launched in beta that are Yeah. Basically it's for free and it is trigger dot dev for free. Okay.
24:29But it's part of the the Vercel ecosystem if you're in in into that ecosystem. But they're kind of these it's a way to monitor, retry, and have these long working tasks. Because as you know, with edge functions, that's been the challenge.
24:42Right? Like, you don't really get a whole hell of a lot of time and things get stale. So anyway, this is nice.
24:50I like it. It's yeah. You're right, though.
24:52It is an extra expense. But I think you can it is relatively inexpensive for this type of task. Alright.
24:59So back to the clusters. So now what we've we've done is we've run an algorithm on top of that.
25:06We've got all the vector embeds, and we're starting to build clusters. So as you can see here, this is today.
25:12I don't know that everything's caught up. It looks like we've enriched oh, yeah. We're close.
25:162284 of 2288 stories have been so 99.8% have been enriched in the last twenty four hours, meaning that they've gone through that entire pipeline.
25:25Of that, ones that were more or less failures that we actually had to reach out to Gemini and and get some extended data from them.
25:33353 stories were done there. And you can see here we're starting to get some actual weight.
25:39So a 105 sources reported on this EU investigation into x over Grok generated sexual images.
25:49Nvidia invest 2,000,000,000 into debt ridden core weave. It's 47 stories.
25:55And you're starting to get something that looks kind of like a tech meme ish type feed. And then obviously you can slice and dice this in a variety of different ways. But this is where it gets even crazier.
26:04So I promised you crazy stuff. This is where we go down the rabbit hole. So let's just take this idea of CoreWeave and NVIDIA investing, uh, 2,000,000,000.
26:15So we'll click on that cluster. And now we have 47 stories.
26:20Nine were done via RSS and 38 discovered. So what does that mean? That means that when I see enough signal and I consider that three or more RSS stories talking about the same thing, I then hit search APIs.
26:31You can hit Brave's Brave's search API. I use a Tableau search API and you can say, go out and find me other stories outside of my RSS ecosystem that may match this and bring them in.
26:47So I can expand the scope and see is this a broader story that has more context that I'm just not seeing because of my finite set of RSS articles. Does that make sense?
26:59Absolutely. Okay. So and then I'm looking at there's average distance between articles and similarities.
27:04This is all part of this this algorithm right here that I'm using for the clustering. So first story started four hours ago. Last minutes was thirteen minutes minutes ago.
27:12So then I created something called the gravity engine. And so this is kind of like a a editorial type score, and then it has actually it votes on it on how important this story is generally speaking.
27:25And I rate these in a few different buckets here, like the impact, Another one called this it's gravity, which we can get into.
27:32The confidence ratio that this is pertains to technology and the things that I'm interested in. And then this is all the stories that it's evaluated. And then here's a little matrix here that I've had built out where you can see you have impact and then you have gravity and it's in the high high area here.
27:47Viral potential and how bubble size, like how big this will eventually get and how many people will be talking about it within tech is 60%. Early trend, like a growing trend, is 75%. Intellectual gravity is 86%, and impact is 88%.
28:02And so a total editorial vote of 85%. And then we can get into the the rubric here, which is the so these are the impact dimensions.
28:13So x axis drivers are industry impact. So it's 90%. So the massive $2,000,000,000 investment solidifies CoreWeave's position as a key new cloud provider, intensifying competition with hyper scalers.
28:25Critically, NVIDIA is also launching its Vera CPU as a standalone product directly challenging Intel and AMD in the data center market. Uh, consumer impact, low, 10%.
28:35This is negligible for consumers. So rated that. Actionability.
28:39Do I need to actually do anything here? Is this for a general audience? Can I take action?
28:4430% said some action for investors maybe, but that's kind of it. And then the risk and urgency, like, do I need to act on this?
28:51So this would be like, okay. There's a nine one one iOS patch that you need to apply to your phone or something. Right?
28:57That would be pegged at like 99% risk and urgency. And then the the y drivers here are intellectual gravity drivers, which is how novel is this, which is very important to me because I want high and I'm not I haven't fine tuned this in because it's saying how novel is this for today.
29:15But in reality, what I want is novelty as applied against the longer horizon timeline because one of the things that's really fun is if you go back and you look at the first time, uh, on, uh, Hacker News when Bitcoin was mentioned, everyone was like, ah, this is stupid, blah blah blah.
29:32But if if I would have found that in here, the novelty would have been off the charts. And that's important to me because oftentimes, the largest things that we do in tech and that we see evolve that turn into these blockbuster things over time seem very silly when we first hear about them or so odd or different.
29:53And that's the signal as both a builder and investor that I want to find as early as I possibly can. Right? So that that's that's one that I care about deeply.
30:03Uh, technical depth, um, you know, second order potential, builder relevance. This was important to me on the AI side.
30:11I wanna know how important it like, as a someone that is building in tech, how much should I pay attention to this entertainment value, which didn't have a lot of that. And then I've got some cross cutting signals here, signals to noise ratio, viral potential, and early trend detection as well.
30:28And then I'd added in some other like, I could these judges just sit here and kind of walk through this, which is, you know, what's the PR fluff risk? Like, how much of this is just a PR thing?
30:38Because a lot of this can be you know, we'll see 20 or 30 and I can see this. It's it's really crazy.
30:46You can tell I can detect when something is a paid sponsorship even when sometimes people aren't calling it out as such, is like really scary and illegal because I will see I can detect the similarity and distance difference between the vectors of the published content between the news articles.
31:05And they're all released with plus or minus an hour of each other. And they're all they're all hitting the same major key points. And it's just like AI reworking the paid sponsorship.
31:17And I'm just like, it's pretty effed up. Like, you know and and I'm like starting to find this stuff.
31:23You know? It it's just like, wow. So this is this is what I do at night.
31:28Greg, I'm I'm embarrassed, but this is what I do at night.
31:31So just a question on this whole this whole feature. Because like when I didn't expect it to be this full fledged, like, on point.
31:41Like, is this something that you sketched out on paper and built it? Or are you, like, working with, you know, AI as your cofounder to kinda help you come up with this?
31:51Like, walk me through the product management piece of this.
31:56Yeah. I mean, I I I I'm very much someone that that builds based on gut instinct.
32:02And for me, 99.9% of the features that I create and and don't get me wrong, there's lot of stuff I would cut out here now.
32:10You know? Like, because I would consider like, ah, I didn't like that. I should probably cut that out.
32:13Like, source distribution is a great one. Like, we can get into. But I started was like, okay, let's crawl RSS.
32:21Okay? Let let's just do that. Let's throw images up.
32:23Let's get as much rich data as we can. Let's do a very basic clustering algorithm even before we put in vector embeds. What does that look like?
32:31How does it feel? And then I just was like, well, what is tech me not doing that I I really care about?
32:38And and that led to my personal curiosity around novelty of of objects. How can I detect these trends before they become big? You know, all these things where I was just like, I'm not seeing that anywhere else.
32:51So I should just build it. Right? And so it's one feature at a time.
32:54So what you're seeing here actually is like, I would say each of these things that you're seeing is probably a day or two of just being like, well, let's see if this will work and what it looks like.
33:07And then putting it and then, you know, kind of using to in today's tools, what I would do is I would just, you know, use compound engineering on Claude code and do a a workflow and say, this is the idea that I have. Actually, I probably would well, it depends on how big of idea.
33:25If it's a minor feature tweak, I would just do it that way. If it's something bigger that I needed to flesh out that had technology that I didn't know what the right choice was, then I would use AI as more of a sparring partner on that side.
33:38So I'll give you a great example. There are probably 10 competing clustering algorithms that you can use for news. And hell if I know which one's the best.
33:46Right? And so I actually took the top two that it recommended for what I was trying to do.
33:52And you you don't what you see here at the top of this URL is see clusters v two. And the reason it says v two is because the second one ended up being better than the first one. The first one no longer exists.
34:02And so it's a lot of kinda like just going down that rabbit hole and saying, well, let's let's try these things out. And and sometimes I would immediately realize that was the wrong direction and just actually just uncommit that last whole GitHub repo or that commit and and PR that I spent four hours on and just chunk it and throw it away altogether.
34:25And be like, well, that was four hours lost, but at the same time, I learned something new. And that's that's all of that's all of building. All building is is failure after failure and just and that and failure is awesome because it's just admitting that you've learned something new.
34:39So many people beat themselves up over failure, and I I don't see it that way.
34:45I just see it as like that's failures like it's the best part. That means the next time it's gonna be a little bit better, you know?
34:51So anyway.
34:53I'm really interested in the this whole concept of like, uh, testing products on synthetic audiences. I don't know if you saw this, but, you know, a few weeks ago, I think Toby from Shopify launched a feature where it was like, you launch your ecommerce store, but, like, based on synthetic AI audiences, here's how they would perform.
35:17Here's how your conversion rate would look like. Here's the products that they would click into based on these personas. And I think that's, like, that's sort of the direction we're probably gonna head in.
35:28So before you like publish something, you know, before you publish an ad, you have some certainty that it's gonna work. Before you publish Right.
35:35An article, you know that, you know, people are gonna be clicking into it. Right.
35:40Yeah. I I I think there is so many domains where that makes a lot of sense. And and Toby's obviously just a freaking genius and so brilliant.
35:51What I do that's slightly different and I would I'd be all for that type of thing, but is I build for myself. I think we're entering into this era of of personal software. Right?
36:00Like if there's a workout app that you don't like because the buttons placement doesn't doesn't do it for you or doesn't track one key core metric, like you just build your own. Right?
36:10And like that's going to be the norm in, you know, if it isn't already, it's going to the norm in like six months from now. Right? And then the question is how many people are there like you that also care about said thing?
36:22Right? And so when I'm building this particular slice of the news or the industry, and we haven't gotten into the understanding of who's touching things because I think one of the things that that is done really well in Techmeme is that social touch of like, you know, when Marc Andreessen touches something with a tweet or what you know, how much more credibility and weight does that add to it even more so than Bloomberg or Wall Street Journal or Business Insider writing about something.
36:48Right? So that's yet another thing that needs to be baked in here as well. But then, you know, if I enjoy it, the nice thing about this whole thing is this was, you know, maybe $300 in AI credits or something, you know, to go build this whole thing.
37:03And I could stand this up, cash the crap out of it, meaning like so that it's performant, and and just put it out there and everyone would be like, okay. If I'm also into the geeky things that Kevin is, I will use this too.
37:17And that might be a thousand people, might be a 100,000 people. No one knows. But also at the same time, like that's okay.
37:23Like if you if you have 500 people that really love what you've created, I get a lot of value and joy out of that.
37:31And it's like we we think in numbers now on the Internet in terms of, you know, millions and billions of people. But in reality, if you like went outside your house and there are 500 people standing like cheering you on, you'd be like, I'm the biggest rock star in the freaking world.
37:46Right? So we lose this perspective of what it means, what success means.
37:51And and so I just I hope that we can all agree and and just realize it doesn't have to we don't have to swing for the fences. So yeah.
37:59So I I just, you know, that's I think what why I'm so excited about this is this democratizing code quality or coding for everyone.
38:06And I was a computer science, uh, major. I dropped out, and I didn't know why.
38:13And I was really slow, and I could understand the core concepts, but I didn't know why I couldn't just do it as fast as my my my everyone I was working with in my in my class. And just six months ago, I found out that I have something called aphantasia, which is this inability to have a mind's eye.
38:32So when, like, people close their eyes, they and they say picture an apple, um, or, like, you know, the famous one of, like, when you're trying to go to sleep, like like sheep jumping over a fence or some shit like that. I always thought they were joking.
38:43You know? I didn't know that you could actually close your eyes and envision things. And so because of that, you know, things like how to handle proper syntax and on on in code and like all the things I was trying to beat into my brain, the retention just wasn't there.
38:58Somehow, the core concepts stick with me and the and the creativity, uh, I've always had that in abundance, which is great. But, uh, but yeah.
39:08And now, like, AI will fill in the deficiencies wherever they are for you, which is just beautiful.
39:15So I also dropped out of CS school. I only have three classes left. And I was the same way in the sense that for me, I wasn't I couldn't get the last 10%.
39:28So like I couldn't get the code to compile because there was like I had I had 92% of it there, but there was, you know, an integer missing here, a variable missing there. And what's cool is nowadays, you know, of course it's nice to know that stuff.
39:43But if you're just trying to get something out the door and to get feedback from people, you know, let Claude Code figure that out for you. Yeah.
39:51Exactly.
39:52And the other thing too I I I think that is lost on a lot of people that I see on social media around what, you know, quote unquote vibe coding means is they say, well, great. And this has been the complaint for a while.
40:04Vibe coding is buggy. It's not performing. It'll fall over underweight, blah blah.
40:08I would argue those are great problems to have. The hardest thing to do is to find something that somebody actually wants to use.
40:17Right? Like, that's the hard problem.
40:21If I have something that I've vibe coded and I launch it and if it crashes on the weight of 50,000 people beating my door down because it's the the next best thing, I guarantee you I can find new engineers to work on that and scale it. Right?
40:35And so I I don't think that should be a reason why we we kind of like I like it because you get more shots on goal.
40:43Right? Like, I don't have to look at the code. It's not because like you, yes, I can jump into a component in TypeScript and and be like, okay, I I kinda see what it's doing here.
40:52You know, can we can we can do that. It's slow, but I can do that. But that's not the point.
40:57I don't care right now. I'd rather see actual humans using it and saying, yeah, Kevin, this is so cool. I want you to formalize this a little bit more, make sure it does scale.
41:06And if if that comes in the form of usage, then I'll find the find the right engineers to to take my my kind of scribble code and and make it real and performant. And and honestly, compound engineering has already been amazing and that it's it's finding a bunch of stuff and and making things more performant for me on the fly.
41:23So for this particular project, you're gonna put it out like success. I'm just trying to think like success. Success looks like what?
41:31Well, it doesn't look like anything. I might never launch this. Like, I I might put together a little one pager that's just like the best AI stories and shown by the things that I care about like novelty and impact or whatever may be or, you know, there's just I I realized that you've got Product Hunt, which is kind of people that have already launched things.
41:51You've got, you know, Techmeme, which is fantastic at overarching big news, you know, Coreweave, $2,000,000,000.
42:00Like, that's a Techmeme story all day long. Right? And then you've got x, which is just a lot of stuff that's coming at us so hot and heavy.
42:11And it's just like, okay, well, where do I decide to spend my time? Like, we were just talking about this before the the podcast started.
42:18You're like, oh, if you play with this, you play with that. And I'm like, ah. You know?
42:21Because it's like you get so there's so much coming at you. I want this thing to eventually if it ever sees the light of day, what I wanted to do is to say, Kevin, this is important to you because it maps to you.
42:35These very important people have touched it and said it's worth your time. And it's past that threshold to where now you should go install it, play, have fun, and learn about it.
42:47Because otherwise, I'm just gonna be and and sadly because of my ADHD, I'm gonna be bouncing around too much to even get anything done. So that's that's my hope. But, yeah, that's so you can see there's a lot of other features we can get in here that we don't have to.
43:00But it's, you know, it it this is me playing to see because a perfect product here actually would be to cut 90% of these features and just find the 10% that really means something to me and a lot of people and launch it as a standalone single page website.
43:18Right? And that's what I'll but this is the messiness of it all. And I I just wanna show people that like for me what I do is I put everything on the table.
43:29All the stuff on the table, all the stuff that I would never show anybody like the the the distance and similarity scores between two stores. Maybe I would show that. But get it all out there and then decide to cut.
43:41And and go in there with that kinda director or editor's cut and start making and start trimming, trimming, trimming, trimming down to down to where you eventually get something that's really usable and useful to to folks.
43:52One feature which I would love to see on something like this would be I'm I'm just thinking out loud. Like, if I was a PM on this product, I would be like, what is the mechanic to bring people back?
44:04So if you go to go to ideabrowser.com. Mhmm. So I built this.
44:11It started off as like a lead magnet actually.
44:17Yeah. I've I've I've I've seen this by the way. This is like a very famous thing now that you've built.
44:22And
44:24so basically, the idea was the like the first the lead magnet started off as here's a database of 30 ideas I would build today.
44:33Yeah. And then, you know, it was like, put in your email to get the access to the database.
44:39And then it was like, okay, how about what we do to make it more fun is almost like you remember Groupon? Let's do like, instead of a product today, an idea of the day. Mhmm.
44:49And then the mechanic is the email every single day. You know, we get a 50% open rate, people open up the email, and it's grown, you know, very very fast through that way.
45:01So I wonder, you know, if I'm building or you're building nile you know, nylon, if it does see the light of day, it's like what is the mechanic to get people back to the website?
45:12Yeah. I mean, in in my mind, the at least for for for news in general, it is relevance to the to the end user.
45:23You know, if if if it is discovering stuff that you are missing or you have overlooked and it's saving you time and energy because it's saying, you know, it's presented in a visual way. Like, I'll give you an example.
45:35If you were on a conference call today and you had, you know, this call like this like the some leading minds in tech of Sam Altman and Marc Andreessen and like a handful of other people in AI said, hey, um, you know, Greg, I think you should go check this out and play with it.
45:53There's a good chance by that afternoon you would be like installing it and and messing around with it. Right? And so visually, I would have to say there here are a thousand signals that came in today.
46:05How can I show you the five things that have launched or that are in beta, that are in GitHub, that are worth your time and map to your interests as well? Right?
46:15So I would probably go in and pull your last 100 or 500 x posts and and look at how you interact and and and create, you know, vector representations of who you are as an individual.
46:28And then also try and get a little a little bit more custom so that it maps to you. So if you're very much into robotics, you would see a bunch of amazing kind of robotics information being presented to you across a variety of different fronts.
46:41So, you know, you would see it both in terms of things that are being talked about on x and also things that are being talked about on Product Hunt or things that are you'd be talked on a hacker news or, you know, Reddit or any number of sources of the new dig or you name it. So I don't know.
46:55I I I really don't I I think at the end of the day for a consumer app, you know, it has to come down to am I finding something on useful from this site or service that I don't see anywhere else? And is it helping me save time and energy?
47:12And the answer may be no. And then guess what? It's it cost me $500 and I flush it down the drain and onto the next thing.
47:18You know? Yeah. Like, I got I've another one I could show you if we have time.
47:22Yeah. We don't have time. Do we have two minutes?
47:24Yeah. Let's do it. Okay.
47:25So this one is is like December 15. So December 15 of so twelve years ago, I posted this idea that you could have a blog where you can actually see the person in the background in real time, but it's blurred out.
47:46But it gives you a sense of kind of presence that they were actually there. And I was I'm still kinda enamored with this idea, especially because we're entering into a world like there there I am like inputting in information or typing as I'm typing.
48:02But we're we're we're entering into this world where we don't even know if there's another human on the other side of it. I think that's just gonna get worse and worse over time.
48:09And so, you know, I, like, full on built this in, uh, Claude code, and now I have it, you know, completely running.
48:19And I can I can just show it to you real quick? Let's see here. Let's pull it up.
48:25While you're pulling that up, it just it strikes me that there's a bunch of good ideas from, for example, twelve years ago that could be recreated today.
48:34Oh, a 100%. Right? Yeah.
48:36I mean, were things that were just either so much of this is just right idea at the right time. You know?
48:45And and some of it was not technically possible. These are the crazy ideas that we had.
48:50Or some of it is even richer now that we can bolt on AI and make it, you know, cooler and different in some unique way. So there are a lot of things that I feel will see the light of day again, but slightly modified because it the cost to do so is is is next to nothing, you know, which is which is great.
49:10So what wasn't possible back then was real time video compression in the browser.
49:17And the reason I say that is the prototype I built twelve years ago, the issue would have been that you uploaded raw video to the site and then you blurred it after the fact, and which meant a user could go trim away the CSS and actually see that person in a very awkward kind of position or whatever and like maybe an environment where they were like looking for a little bit of blurred privacy.
49:39And so now this is like all done in real time. So what you're seeing now is that it's actually recording me.
49:45It's like like, hello world. This is a test.
49:51And then I could just kinda move my arms around like this a little bit, Get a little movement and and please like this is the v point o one alpha of this whole thing. And then hit submit and it's doing some horrible broken math here.
50:06But you see that that's actually me in the background and that's not me now. That is the blurred version of me. And it's a horrible interface.
50:14I would never release this interface. But see that was me when I was waving my hands a second ago. But I could do all of the compression now on client side so that I do preserve privacy and put this up in the background.
50:27And so you can imagine these little slivers and this little visibility into people's world as they're kind of blogging. And in my head, I'm like, okay.
50:36Well, maybe this should just be my blog and I'll release this and I'll just open source it and give it away. And there doesn't need to be a business model because this is just fun.
50:46You know? Totally. So like that's that's what I love about the the time we're living in, man.
50:50It's the best. You just have fun.
50:54You can just you could just put out things and and sometimes though, it's sort of interesting. It's like when when there's not when you just put up projects for fun, somehow, I don't know why, but they end up being the the the the projects that could end up becoming the biggest businesses.
51:13You are 1000% right. It is the weirdest thing that I do not know how to explain.
51:20Like when I when I did dig back in 2004, it was just like, I just wanna see if people can vote on things and what it looks like when the best stuff hits the homepage. And then, you know, a year and a half later, there's 38,000,000 people a month using the site.
51:32When I made Zero the intermittent fasting app, I was like, I just want a way to track my fast and, like, have a silly little timer. And then, you know, the company is doing double digits millions of dollars in revenue off of a little, like, timer app with some other content on there.
51:46And I'm just like, how is this even these were just for fun. These were you know?
51:51And it it's it's very it's very strange how that works.
51:55So before we wrap up, you're you're working on DIG and you're kinda incubating projects.
52:02Like, I'm listening to you and I'm like, how can I work with Kevin? I'm sure people are listening to this being like, this this sounds cool.
52:09Like, how could people support, get to work with you?
52:14Yeah. I think there's a couple different ways.
52:16So I appreciate you saying that. There's I have a I'm sitting in an office right now that's completely empty.
52:22And so I have an incubator here in LA, a studio that I'm gonna kinda open up. And there's just gonna be free desks for people to come in that are jamming on really cool stuff. So if, you know, at reply me or DM me on x if you're you're building really cool AI stuff and you even if you're just coming through LA and just wanna come in and jam.
52:41And it it and the point is like, let's just compare notes and talk about what's cool over lunch. And like, you know, if you need a office or you need a room to take a call, you got it for an hour. Like, no big deal.
52:52Right? No fees or any type of like, I don't wanna charge for desks or anything like that. So I'm gonna surround myself with those types of people here in in Venice, out in LA, and I wanna do that.
53:04Um, so please, like, hit me up and let's let's figure out a way to connect if you're building really cool stuff. And then, you know, I'm a venture capitalist still over at True Ventures. And part of what I think I think VC is gonna evolve dramatically over the next couple of years because I I believe that people don't need to raise capital.
53:21And oftentimes, most ideas, especially like just great lifestyle businesses that get to, you know, $2.04, $510,000,000 in revenue, like, own that shit a 100%. Don't sell it to VCs.
53:31I'm not supposed to be saying that because I'm a VC. And so but like, don't do it.
53:36And so but that's why I hope people will realize that, like, that that's why I there is a time and place for VC.
53:46Like, when you really get to scale and you're like, damn, I need I need a couple million bucks to hire because the growth is so outrageous. You know? Like and and that's that's kind of what I do.
53:56And that's why I wanna play also. I think VCs the the era of VCs just being these like MBAs that sit there and like try to tell you how to run your business, I think is so boring to me.
54:06I want a VC that is playing, that's building alongside me, that's like pressure testing my ideas, that's a thought partner on this stuff.
54:15And so that's kind of like why I don't know what you wanna call I hate the even word venture capital. I think it's like evil.
54:21It feels evil these days. I just if I if if someone does first of all, I'd try to talk everyone out of taking money. And if they do need money, especially hardware companies need a lot of money, then they should find someone that's also building and that's build stuff at scale, not just because somebody has a a great a great pedigree or, you know
54:42yeah. Does that make any sense at all? It does.
54:45And and I just wanna add, like, I'm the first person to be like, you know, people listen to me on the pod know that I think that you should your MVP, your business, you know, you probably don't need venture to start.
54:58And there's some businesses that shouldn't raise venture, but I also think that in this world where you can build software, MVP it, if you wanna build hardware, you're like, you build this cool piece of software, it starts to take off and you're like, know, oh, there's this logical extension to hardware.
55:15You're probably gonna need to raise money. Yeah. A 100%.
55:18We we invest in a company called Sandbar that's doing this AI ring. I don't know if you've seen the prototypes for it. Yeah.
55:23But it's it's really cool. And, you know, they I can't remember how much you put in. It was close to 10,000,000 or something like that.
55:30But it turns out to do the tooling, to do to get to go and, you know, get that to scale, like, that's those are those are real dollars still required to to pull that off. Totally.
55:41Kevin, it's been an absolute treat having you on the pod. I hope you come back on again
55:48and show and tell more because you're up to some really cool stuff. You're a legend, and I'm gonna take you up on that. Next time I'm in LA, I'm gonna Oh, seriously.
55:55Walk through. Yeah. Done.
55:57Like, you you got an office here, so come come hang, and and thanks for having me on. I appreciate it. And thanks for all the work you've been doing in this space, man.
56:03It's like, I I know we don't talk that often except for like the random DMs and stuff, but I see you all over x and the the content and stuff that you're putting out. It's absolutely fantastic.
56:13Thanks, man. Just trying to create some signal out in the world of noise.
56:17It's great. Well, maybe the nylon app will identify that and put it in front of more people. Exactly.
56:21That's what I'm hoping. Alright. Take care, Kevin.
56:24Take care.
The Hook
The bait, then the rug-pull.
Kevin Rose has built and sold things at internet scale — Digg at 38 million monthly users, Zero at double-digit millions in revenue — and now he's spending his nights vibe-coding a personal Techmeme clone tuned for AI velocity. The pitch is straightforward: can one person, with today's models and a couple of weekends, produce a news intelligence tool that rivals what editorial teams take years to build?
How they asked for the click.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
