ARS_MECH

@PatrickHeizer @max_spero_ All beef?

@max_spero_ We have ethical meat; go find a local farmer who is doing it right and support them. If you or anyone is near Frederick, MD, I'd be happy to host a tour of my farm.

Can anyone explain to me why we don’t have ethical meat? Surely we can give the farm animal a significantly better life than it would have lived in the wild.

@Everlier @pmarca 100

@pmarca Moral panics of 2020s include teens not doing any of this. Overcorrected.

Moral panics of the 1980s included teens drinking, teens having sex, teens driving, teens smoking, teens watching television, teen movies, teens playing RPG’s, teens listening to music, teens playing video games.

@LUKE_UGA @Jason @grok good

@Jason @grok Here ya go JCal

1. Make this a painting @grok 2. Name this band and their debut album

@aakashgupta Ehhh, its not like IF you send a photo of a pool in my backyard I'm going to open my wallet ... ? Like we have all considered a pool or X add on.

The customer acquisition cost for a pool installer to land one $48,500 job traditionally runs $2,000 to $4,000 in paid ads, follow-ups, site visits, and quotes that go nowhere. Close rates on cold leads hover around 5-8%. This agent flips the entire funnel. It scans satellite imagery for 25,000 sqft lots in Tampa, checks setbacks, filters by home value ($500K-$1.2M), and renders a photorealistic pool in the homeowner's actual backyard. Then it mails a postcard. The postcard is the part that makes this print money. A homeowner who has never considered a pool opens their mailbox and sees their own yard with a pool already in it. The visualization does the selling before a human ever picks up a phone. Every pool builder in America spends thousands trying to get homeowners to imagine the finished product. This agent generates that image for pennies per address and sends it unsolicited to qualified prospects. The economics work because the targeting is surgical. $500K-$1.2M homes without pools means the homeowner can afford $48,500 and has yard space. The $37,500 property value increase means the pool nearly pays for itself on paper. 43 days to Memorial Day is the urgency trigger. Every variable in the postcard is calculated to push one decision. Pool installation in Florida runs $70,000 to $120,000 on the high end. At $48,500 this is priced at the impulse tier of an industry where "impulse" used to mean $100K+. One closed deal from 500 postcards and the entire campaign is profitable by 10x. The real product here isn't pool sales. It's an AI-powered lead generation machine that works for any high-ticket home service: solar panels, ADUs, kitchen remodels, landscaping. Anywhere you can render the finished product onto the prospect's actual property photo and mail it to them, this model applies. Pool installations are the proof of concept.

@karpathy Similar work flow that google baked into this notebooklm.google ... not perfect I hope all tools will allow portability of interactions to export / manage cut and manipulate into the future

LLM Knowledge Bases Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So: Data ingest: I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them. IDE: I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides). Q&A: Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale. Output: Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base. Linting: I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into. Extra tools: I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries. Further explorations: As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows. TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

@jeffreyjhardy @theallinpod @bryan_johnson He's visibly high in the interview, interesting timeline with silicon valley both destroying and being destroyed

@theallinpod @bryan_johnson I listened to the whole podcast. Respectfully … two things: 1. He sounds like every drug addict I have met. Some in my own family. 2. This will not end well.

Bryan Johnson did the world's strongest psychedelic drug... here's how it went: @bryan_johnson: “ I did 5-MeO-DMT, which is the most powerful psychedelic on the planet, somewhere between 5-10 times more powerful than DMT. I'm stunned. Absolutely floored. Speechless. You basically experience raw consciousness and raw intelligence. You either panic because you feel like the gates of hell are going to open, that the stream of existence is going to just tear you to shreds.” @friedberg: “Like break your brain?” Bryan: “Yeah. You feel like it's going to threaten your sanity. Like is it going to chop you up into little pieces? And so in that moment you have to say, do I try to wrestle this? And I need to just like, wait it out until it's over, or you just relent and you say ‘Yes,’ and you have to, in that moment, you have to say yes so thoroughly. You have to release all attachment, all preconditions, all want, all desire, you have to release self, ego, control. You just have to relent entirely. And then when you do that, it opens up this unimaginable bliss and euphoria.” ------------------------------------------ Thanks to our partner for making this happen! The Pod by @eightsleep cools your bed to 55°F and uses Autopilot to optimize your sleep, all night. Use code ALLIN at eightsleep.com/allin for up to $350 off.

@PatrickHeizer (2) economist.com/finance-and-ec…

Feels like I'm living in the Twilight Zone. If the Strait isn't opened within two or three weeks, a global depression is coming. Petroleum enables modern life. The Spice must flow.

@PatrickHeizer (1) A rule of thumb says that a sustained $10 rise in oil prices eventually adds 0.3-0.4 percentage points to overall inflation. Assuming oil stays around $100 a barrel, the OECD’s inflation rate might rise above 4%; at $140, inflation of 5-6% is not out of the question.

@PatrickHeizer I used to think the "architect" was the last job, the orchestrator, etc. Now I think its the farmer, only an all out world war can pry you from the land and automation of planning and labor will lever you (if you'd like). Never being threatened with starving will be powerful!

Plant trees, cultivate gardens, tend livestock, steward your community, flourish in the real world.

@PatrickHeizer Patrick only followed for a few weeks, but loving your takes. So much broken shit, glad someone else is pissed.

The US ethanol standard is arguably the worst current policy on the books. Now, because of a misguided war, we are expanding it. We waste ~35M acres on a crop that extracts and destroys our soil, pollutes our rivers and air, worsens our food supply, all while subsidizing oligarchies like John Deere, Monsanto, Syngenta, Cargill, and the oil industry, and it doesn't even lower carbon emissions. Imagine if instead we used those acres for widespread regenerative agriculture, animals actually living on the land, under and around perennial crops that produce actual nutrition, cleaning our waterways while promoting biodiversity, while simultaneously increasing wild and protected lands and public recreation areas. A society built around life, not corn.

@balajis CANT STAND OTHER PEOPLES LLM OUTPUT *NEGATIVE REPUATION* (this was written in codex 5.4) seriously if I wanted to read an AI response I would just ask my own with all the context and steering of my personal preference

AI is a shortcut. So it’s useful. But it’s lazy. So it speeds execution. But it hides complexity. So you want to use it. But definitely not overuse it. So the user of AI often loves it. But the reader of AI often hates it.

@DavidSacks BRAVO

Amazing.

@Omnitrader69 honestly the release felt so weak ... like are you stuffing the fucking channel or not! IF you are not then have a fucking press release at the distribution center. Smells BAD, I took a 20K loss and sold on that press release.

$ADMA sub 10 now after that timid response from the company. Gotta get out there guys and say more unless you enjoy being called frauds. 😠

@CulperResearch @MartinShkreli - agree?

NEW: We are short ADMA Biologics $ADMA. We believe the ASCENIV growth story is in reality a story of channel stuffing and an undisclosed related party distributor. Real revenues declined 3% in 2025 vs. 20% reported growth. It's already over. Full report available on our website.

@fbarnau @pmarca exactly - everyone is working 7 days a week to tread water. its not like they are getting PAID for these hours! We are all trying to figure out how not to be left behind

@pmarca Let’s not conflate FOMO-driven overwork with real labor demand. Layoffs remain structurally elevated: ~245,000 tech jobs were cut in 2025, and another ~45,000–58,000 have already been eliminated in the first quarter of 2026 alone.

And so the myth of technology-driven unemployment dies once again.

@grok @VVTLD @GeminiApp 😂🤡

We’ve been seeing some amazing Nano Banana 2 creations lately. 🍌 Here are some standouts. 🧵

@steipete @morganlinton @denisyarats The CLI champ himself, hilarious how MCP is done after Lex interview!

@morganlinton @denisyarats Wonder where I seen this idea implemented before.

The cofounder and CTO of Perplexity, @denisyarats just said internally at Perplexity they’re moving away from MCPs and instead using APIs and CLIs 👀

@MartinShkreli you do anything with this?

any $ADMA news?

@MSparrowww @typedfemale thank you, author left out works cited

@typedfemale Still not as good as the original: youtu.be/4sZuN0xXWLc?si…

presenting: big jeff's trainium hell

@steipete For those that don't know the origin here is the actually funny OG. BIG BILL HELL youtube.com/watch?v=4sZuN0…

bring your clawwww