Key Takeaways
- Bending Spoons-style roll-ups are being framed as a repeatable playbook: buy software and services at 2.3–2.7x ARR, run them with AI-heavy efficiency, and scale toward a $24B model.
- AI stocks are rallying while the weakest US corporate credit spreads hit 12.52%, the widest in nearly four years — two very different markets.
- S&P 500 concentration is at a record: the top five names (Nvidia, Apple, Microsoft, Google, Amazon) exceed 28% of the index, while the equal-weight ratio is at its lowest since November 2002.
- Housing is effectively frozen: over 1.16M homes for sale (highest since 2019), pending sales at the lowest since April 2020, and sellers outnumber buyers by more than 560,000.
- AI-generated music and video projects — "Slopcore" — are being built with 55,000 lines of code, 282 generated clips, and tools like ElevenLabs Music and Grok Imagine.
- A $5,000 bounty submission lets users clip paragraphs, video, or podcast segments from any page, add commentary, and publish with linked sources via a Chrome sidebar.
- Drone delivery is being described as "the future is here" with 18 restaurant options, satellite-map drop-point selection, ETA countdowns, and quiet, precise drops.
1. AI Business Models and the Roll-Up Playbook
- Jason amplified Bending Spoons data showing a $24B market cap, 8x ARR trading multiple, and 2.3–2.7x ARR acquisitions of B2B companies. The cited figures include Q2 revenue growth of 126% but only 3% organic growth, EPS growth of 167%, R&D plus sales and marketing at just 18.6% of revenue, fewer than 700 core team members, over $4M revenue per employee, AI writing more than 90% of code, and customers paying about 10% more for AI features. Jason's view is that ten similar companies will emerge, and that consolidating software and services assets to run them with "superintelligent efficiency" is an excellent business. — via 1
- Marc Andreessen repeatedly shared a creator's "Slopcore" AI music and video project. One post describes a four-day Claude conversation producing a "Berlin rave meets opera" piece with 55,000 lines of code, 282 generated clips, and 99 selected, with music from ElevenLabs Music and AI-generated visuals; the creator calls it "human-made Slopcore." Another, "Mandelbrot Brainrot," credits Caesar with concept, lyrics, and AI art, uses ElevenLabs Music, and thanks @boneGPT for building Slop Cannon. Additional posts mention Grok Imagine and a "give a prompt and go to sleep" workflow, while a repost says "Fire the Cannon and bring to life any world you can conjure up" and calls it "the golden age." "Slop Caesar" is described as a roughly six-minute piece about "shrimp ambition completely out of control," with a note to watch with sound on. Tags include #slopcore and #awesomecore. — via 1 2 3 4 5 6 7
- Jason promoted a $5,000 bounty "annotated" submission that lets users clip paragraphs, video, or podcast segments from any page, add opinions, and publish with linked sources, supported by a Chrome sidebar and a live site. He also ran a caption contest with a $420 X Money prize and, separately, said he started a trend after a "randomly pick one reply in 24 hours for $69.42" giveaway, reposting a thank-you that said the money would go to charity. — via 1 2 3 4
2. Market Structure, Credit, and Concentration
- Charlie Bilello highlighted that AI-related stocks are rising while credit spreads for the weakest US companies have widened to 12.52%, the widest in nearly four years, describing two very different markets. He also noted that since 1928, the S&P 500's average annual return during rising Treasury yield periods is nearly identical to falling-yield periods, challenging the common fear of rising rates. A repost noted the top five S&P 500 weights — Nvidia, Apple, Microsoft, Google, and Amazon — now exceed 28%, the highest concentration on record. — via 1 2 3 4 5
- Bilello contrasted 2020, when the 10-year Treasury yield was below 1% and real yields were negative, with 2026, when the nominal yield of 5.31% is the highest since 2002 and the real yield is the highest since 2008. He noted the S&P 500 was down 7% year-to-date on March 30 — one of the worst starts ever — before a sharp rebound took it to +15% for the year, more than double the average gain of +6.8% over the same period. The equal-weight S&P 500 ratio to the cap-weighted index fell to its lowest since November 2002, erasing more than 23 years of relative outperformance as a few mega-caps continued to dominate. A repost declared TINA dead: 30-year TIPS yield 3.37% above inflation, the highest real yield in over 20 years, giving investors an alternative to stocks. Another repost listed price-to-sales (TTM): SpaceX 103, Palantir 87, CrowdStrike 52, AMD 24, Broadcom 20, Nvidia 19, Tesla 13, Microsoft 12, Apple 11, Google 10, Meta 8, Amazon 4, and the S&P 500 at 3.5. — via 1 2 3 4 5 6 7 8
- Liz Ann Sonders noted that profitable Russell 2000 constituents are consistently outperforming unprofitable ones, reversing last year's pattern when unprofitable stocks beat profitable ones by 2-to-1. The data, from Bloomberg as of October 9, 2026, defines profitable companies as those with positive trailing 12-month EPS; the author reminds that indexes cannot be invested in directly and past performance does not guarantee future results. Sonders also shared charts on market breadth, index drawdowns and hit rates, Tensor10 tables, rolling correlations between bond yields and stock prices, a basket of stocks, and sector/index performance for Friday, month-to-date, and year-to-date. She previewed a CNBC PowerLunch appearance with @SullyCNBC at about 2:15 p.m. ET that day. — via 1 2 3 4 5 6 7 8
- Bilello also flagged a frozen housing market: over 1.16 million homes for sale, the highest since 2019; pending home sales at the lowest since April 2020; and sellers outnumbering buyers by more than 560,000 (58%), the widest gap since data began in 2013. — via 1 2 3
3. Infrastructure, Delivery, and Macro Signals
- Jason cited a drone delivery experience as "the future is here," naming Zipline, ICON3DTech, micro1_ai, goautolane, Hello_GoAI, and AskTaxGPT. The cited post describes choosing from 18 restaurants, selecting and paying for a meal, picking a drop point on a satellite map, receiving an ETA and countdown, and a quiet, precise drone drop. — via 1
- Meb Faber shared Goldman Sachs data showing China accounts for 77% of global token usage but only 22% of token spending. He also relayed Jim Grant's explanation that he chose "observer" rather than "prophet" as his 12-page journal's tagline not out of modesty, but because he cannot predict interest rates and no one can. — via 1 2
- Lyn Alden, a structural bond bear, said the recent bond bear-market surge was not surprising: the yield curve remains flat and nothing has broken, which is almost expected. But as more people discuss it, she may turn neutral on bonds. She also recommended Meb Faber's charts and history book as valuable physical material for macro and market history readers, called Claude's self-"Rick Roll" "Brilliant," and reposted a reminder that she will speak at Philly Bitcoin Jawn about macro, her new book (with a signing afterward), Orange Juice, and current events. A repost declared "gold bulls — this is your moment." — via 1 2 3 4 5
- Clifford Asness reposted skepticism about a Bitcoin-related strategy: earning 3.5% on $3B in cash to offset 12% borrowing costs, mocked as "arbitrage." The repost also questioned why a "BTC Monetization Program" is needed if BTC itself is money, and asked whether someone is shorting $STRC. Asness's own post said, "One, two, three, four, I declare an energy war truce," a phrase that also appeared in a repost. Other reposts said he likes discovering he was just agreeing with Thomas Sowell, and that Putin does not care about Palestinians, like his Soviet predecessors, and uses them and their cause only as a tool to attack the West. — via 1 2 3 4 5 6 7
- Jason also posted "Kappabashi 🔪" with a link, reposted Upfront Heating & Air Conditioning content, and said he let a bot randomly send X Money to repliers, calling it "randomly giving money to new followers" and describing X Money as an "infinite money glitch." — via 1 2 3 4
