Offshore
Photo
Dimitry Nakhla | Babylon Capitalยฎ
RT @DimitryNakhla: Nick Sleep is one of the greatest investors, generating 20.8% CAGR from 2001 to 2013 ๐
Nick ran a concentrated fund, with $AMZN & $COST stock each accounting for >20% of the Nomad Partnership at various points ๐ธ
Hereโs the secret to his incredible track record ๐งต https://t.co/9mDupZqv57
tweet
RT @DimitryNakhla: Nick Sleep is one of the greatest investors, generating 20.8% CAGR from 2001 to 2013 ๐
Nick ran a concentrated fund, with $AMZN & $COST stock each accounting for >20% of the Nomad Partnership at various points ๐ธ
Hereโs the secret to his incredible track record ๐งต https://t.co/9mDupZqv57
tweet
Dimitry Nakhla | Babylon Capitalยฎ
5 High-Quality Stocks With Good CAGR Potential Assuming Reasonable Multiples ๐ต
๐ฆ Amazon $AMZN
โข2028E EPS: $11.01
โขMultiple: 29x
โขCAGR: +12%
๐ฐ S&P Global $SPGI
โข2028E EPS: $24.41
โขMultiple: 28x
โขCAGR: +13%
๐ฆ Fair Isaac $FICO
โข2028E EPS: $63.60
โขMultiple: 36x
โขCAGR: +13%
โ๏ธ Booking Holdings $BKNG
โข2028E EPS: $344.04
โขMultiple: 23x
โขCAGR: +15%
๐ซฑ๐ผโ๐ซฒ๐ป MercadoLibre $MELI
โข2028E EPS: $123.53
โขMultiple: 30x
โขCAGR: +18%
_______
*Estimates can change
tweet
5 High-Quality Stocks With Good CAGR Potential Assuming Reasonable Multiples ๐ต
๐ฆ Amazon $AMZN
โข2028E EPS: $11.01
โขMultiple: 29x
โขCAGR: +12%
๐ฐ S&P Global $SPGI
โข2028E EPS: $24.41
โขMultiple: 28x
โขCAGR: +13%
๐ฆ Fair Isaac $FICO
โข2028E EPS: $63.60
โขMultiple: 36x
โขCAGR: +13%
โ๏ธ Booking Holdings $BKNG
โข2028E EPS: $344.04
โขMultiple: 23x
โขCAGR: +15%
๐ซฑ๐ผโ๐ซฒ๐ป MercadoLibre $MELI
โข2028E EPS: $123.53
โขMultiple: 30x
โขCAGR: +18%
_______
*Estimates can change
tweet
Offshore
Photo
ToffCap
I wanted to publish a larger update note, but I ended up writing about a French micro-cap, ๐ซ๐ท HighCo $HCO.
This one has been on the screen for a looong time, but only recently made the cut:
- A sleepy French promo outfit finally finding its way up operationally.
- The domestic engine (France) is back to double-digit growth.
- Two pretty nice bolt-ons, tipping the business toward scalable, recurring business.
- Guidance is actually going up!
- c. 5x forward ebitda, dd fcf yield and >6% dividend yield for a business thatโs starting to look like a platform, not an agency.
Let me know what you think @ToffCap (no paywall) ๐
tweet
I wanted to publish a larger update note, but I ended up writing about a French micro-cap, ๐ซ๐ท HighCo $HCO.
This one has been on the screen for a looong time, but only recently made the cut:
- A sleepy French promo outfit finally finding its way up operationally.
- The domestic engine (France) is back to double-digit growth.
- Two pretty nice bolt-ons, tipping the business toward scalable, recurring business.
- Guidance is actually going up!
- c. 5x forward ebitda, dd fcf yield and >6% dividend yield for a business thatโs starting to look like a platform, not an agency.
Let me know what you think @ToffCap (no paywall) ๐
tweet
Offshore
Photo
Dimitry Nakhla | Babylon Capitalยฎ
A quality valuation analysis on $APP ๐ง๐ฝโโ๏ธ
โขNTM P/E Ratio: 42x
โขYTD-Mean: 44x
โขNTM FCF Yield: 2.3%
โขYTD-Mean: 2.5%
As you can see, $APP appears to be trading near fair value
Going forward, investors can receive about the same in earnings per share & FCF per share ๐ง ***
Before we get into valuation, letโs take a look at why $APP is a good business
BALANCE SHEETโ
โขCash & Short-Term Inv: $1.19B
โขLong-Term Debt: $3.51B
$APP has a good balance sheet, a BBB S&P Credit Rating, & 11x FFO Interest Coverage
RETURN ON CAPITAL๐โก๏ธโ
โข2021: 3.0%
โข2022: (0.4%)
โข2023: 14.6%
โข2024: 38.9%
โขLTM: 62.5%
RETURN ON EQUITY๐โก๏ธโ
โข2021: 3.6%
โข2022: (9.5%)
โข2023: 22.6%
โข2024: 124.7%
โขLTM: 253.8%
$APP has strong returns on capital, highlighting the financial efficiency of the business
REVENUESโ
โข2021: $2.79B
โข2024: $4.71B
โขCAGR: 19.07%
FREE CASH FLOWโ
โข2021: $360.46M
โข2024: $2.09B
โขCAGR: 79.77%
NORMALIZED EPSโ
โข2021: $1.39
โข2024: $5.69
โขCAGR: 59.96%
SHARE BUYBACKSโ
โข2021 Shares Outstanding: 342.76M
โขLTM Shares Outstanding: 345.41M
MARGINSโ
โขLTM Gross Margins: 78.6%
โขLTM Operating Margins: 51.0%
โขLTM Net Income Margins: 43.8%
***NOW TO VALUATION ๐ง
As stated above, investors can expect to receive about the same in EPS & FCF per share
Using Benjamin Grahamโs 2G rule of thumb, $APP has to grow earnings at a 21% CAGR over the next several years to justify its valuation
Today, analysts anticipate 2026 - 2028 EPS growth over the next few years to be just greater than the (21%) required growth rate:
2025E: $9.26 (62% YoY) *FY Dec
2026E: $14.06 (52% YoY)
2027E: $18.32 (30% YoY)
2028E: $25.39 (38% YoY)
$APP has a good track record of meeting analyst estimates ~2 years out, but letโs assume $APP ends 2028 with $25.39 in EPS & see its CAGR potential assuming different multiples
40x P/E: $1015๐ต โฆ ~18.6% CAGR
38x P/E: $964๐ต โฆ ~16.7% CAGR
36x P/E: $914๐ต โฆ ~14.7% CAGR
34x P/E: $863๐ต โฆ ~12.7% CAGR
32x P/E: $812๐ต โฆ ~10.6% CAGR
As you can see, $APP appears to have attractive return potential if we assume >36x earnings multiple (allowing for ~15% multiple compression)
Growing its monetization of gaming, user acquisition & subscription business, $APP potential catalysts include broadening of its software platform beyond mobile games into enterprise marketing automation, and leveraging its data-rich advertising stack for AI-driven optimization โ both of which could unlock multiple years of EPS expansion
Despite its potential and AXON rollout, $APP is not for the faint-hearted, as its extreme volatility can test even the most resolute shareholdersโ conviction in the business
Additionally, $APP also comes with a high-level of uncertainty
$APP is a good business & appears to be a fair consideration for today at $589๐ต
___
๐๐๐๐๐๐๐๐๐๐โผ๏ธ: ๐๐ก๐ข๐ฌ ๐ข๐ฌ ๐๐๐ ๐๐ง๐ฏ๐๐ฌ๐ญ๐ฆ๐๐ง๐ญ ๐๐๐ฏ๐ข๐๐. ๐๐๐๐ฒ๐ฅ๐จ๐ง ๐๐๐ฉ๐ข๐ญ๐๐ฅยฎ ๐๐ง๐ ๐ข๐ญ๐ฌ ๐ซ๐๐ฉ๐ซ๐๐ฌ๐๐ง๐ญ๐๐ญ๐ข๐ฏ๐๐ฌ ๐ฆ๐๐ฒ ๐ก๐๐ฏ๐ ๐ฉ๐จ๐ฌ๐ข๐ญ๐ข๐จ๐ง๐ฌ ๐ข๐ง ๐ญ๐ก๐ ๐ฌ๐๐๐ฎ๐ซ๐ข๐ญ๐ข๐๐ฌ ๐๐ข๐ฌ๐๐ฎ๐ฌ๐ฌ๐๐ ๐ข๐ง ๐ญ๐ก๐ข๐ฌ ๐ญ๐ฐ๐๐๐ญ.
๐๐ก๐ ๐ข๐ง๐๐จ๐ซ๐ฆ๐๐ญ๐ข๐จ๐ง ๐๐จ๐ง๐ญ๐๐ข๐ง๐๐ ๐ข๐ง ๐ญ๐ก๐ข๐ฌ ๐ญ๐ฐ๐๐๐ญ ๐ข๐ฌ ๐ข๐ง๐ญ๐๐ง๐๐๐ ๐๐จ๐ซ ๐ข๐ง๐๐จ๐ซ๐ฆ๐๐ญ๐ข๐จ๐ง๐๐ฅ ๐ฉ๐ฎ๐ซ๐ฉ๐จ๐ฌ๐๐ฌ ๐จ๐ง๐ฅ๐ฒ ๐๐ง๐ ๐ฌ๐ก๐จ๐ฎ๐ฅ๐ ๐ง๐จ๐ญ ๐๐ ๐๐จ๐ง๐ฌ๐ญ๐ซ๐ฎ๐๐ ๐๐ฌ ๐ข๐ง๐ฏ๐๐ฌ๐ญ๐ฆ๐๐ง๐ญ ๐๐๐ฏ๐ข๐๐ ๐ญ๐จ ๐ฆ๐๐๐ญ ๐ญ๐ก๐ ๐ฌ๐ฉ๐๐๐ข๐๐ข๐ ๐ง๐๐๐๐ฌ ๐จ๐ ๐๐ง๐ฒ ๐ข๐ง๐๐ข๐ฏ๐ข๐๐ฎ๐๐ฅ ๐จ๐ซ ๐ฌ๐ข๐ญ๐ฎ๐๐ญ๐ข๐จ๐ง. ๐๐๐ฌ๐ญ ๐ฉ๐๐ซ๐๐จ๐ซ๐ฆ๐๐ง๐๐ ๐ข๐ฌ ๐ง๐จ ๐ ๐ฎ๐๐ซ๐๐ง๐ญ๐๐ ๐จ๐ ๐๐ฎ๐ญ๐ฎ๐ซ๐ ๐ซ๐๐ฌ๐ฎ๐ฅ๐ญ๐ฌ.
๐๐ง๐๐จ๐ซ๐ฆ๐๐ญ๐ข๐จ๐ง ๐๐จ๐ง๐ญ๐๐ข๐ง๐๐ ๐ข๐ง ๐ญ๐ก๐ข๐ฌ ๐ญ๐ฐ๐๐๐ญ ๐ก๐๐ฌ ๐๐๐๐ง ๐จ๐๐ญ๐๐ข๐ง๐๐ ๐๐ซ๐จ๐ฆ ๐ฌ๐จ๐ฎ๐ซ๐๐๐ฌ ๐๐๐ฅ๐ข๐๐ฏ๐๐ ๐ญ๐จ ๐๐ ๐ซ๐๐ฅ๐ข๐๐๐ฅ๐, ๐๐ฎ๐ญ ๐ข๐ฌ ๐ง๐จ๐ญ ๐ ๐ฎ๐๐ซ๐๐ง๐ญ๐๐๐ ๐๐ฌ ๐ญ๐จ ๐๐จ๐ฆ๐ฉ๐ฅ๐๐ญ๐๐ง๐๐ฌ๐ฌ ๐จ๐ซ ๐๐๐๐ฎ๐ซ๐๐๐ฒ.
tweet
A quality valuation analysis on $APP ๐ง๐ฝโโ๏ธ
โขNTM P/E Ratio: 42x
โขYTD-Mean: 44x
โขNTM FCF Yield: 2.3%
โขYTD-Mean: 2.5%
As you can see, $APP appears to be trading near fair value
Going forward, investors can receive about the same in earnings per share & FCF per share ๐ง ***
Before we get into valuation, letโs take a look at why $APP is a good business
BALANCE SHEETโ
โขCash & Short-Term Inv: $1.19B
โขLong-Term Debt: $3.51B
$APP has a good balance sheet, a BBB S&P Credit Rating, & 11x FFO Interest Coverage
RETURN ON CAPITAL๐โก๏ธโ
โข2021: 3.0%
โข2022: (0.4%)
โข2023: 14.6%
โข2024: 38.9%
โขLTM: 62.5%
RETURN ON EQUITY๐โก๏ธโ
โข2021: 3.6%
โข2022: (9.5%)
โข2023: 22.6%
โข2024: 124.7%
โขLTM: 253.8%
$APP has strong returns on capital, highlighting the financial efficiency of the business
REVENUESโ
โข2021: $2.79B
โข2024: $4.71B
โขCAGR: 19.07%
FREE CASH FLOWโ
โข2021: $360.46M
โข2024: $2.09B
โขCAGR: 79.77%
NORMALIZED EPSโ
โข2021: $1.39
โข2024: $5.69
โขCAGR: 59.96%
SHARE BUYBACKSโ
โข2021 Shares Outstanding: 342.76M
โขLTM Shares Outstanding: 345.41M
MARGINSโ
โขLTM Gross Margins: 78.6%
โขLTM Operating Margins: 51.0%
โขLTM Net Income Margins: 43.8%
***NOW TO VALUATION ๐ง
As stated above, investors can expect to receive about the same in EPS & FCF per share
Using Benjamin Grahamโs 2G rule of thumb, $APP has to grow earnings at a 21% CAGR over the next several years to justify its valuation
Today, analysts anticipate 2026 - 2028 EPS growth over the next few years to be just greater than the (21%) required growth rate:
2025E: $9.26 (62% YoY) *FY Dec
2026E: $14.06 (52% YoY)
2027E: $18.32 (30% YoY)
2028E: $25.39 (38% YoY)
$APP has a good track record of meeting analyst estimates ~2 years out, but letโs assume $APP ends 2028 with $25.39 in EPS & see its CAGR potential assuming different multiples
40x P/E: $1015๐ต โฆ ~18.6% CAGR
38x P/E: $964๐ต โฆ ~16.7% CAGR
36x P/E: $914๐ต โฆ ~14.7% CAGR
34x P/E: $863๐ต โฆ ~12.7% CAGR
32x P/E: $812๐ต โฆ ~10.6% CAGR
As you can see, $APP appears to have attractive return potential if we assume >36x earnings multiple (allowing for ~15% multiple compression)
Growing its monetization of gaming, user acquisition & subscription business, $APP potential catalysts include broadening of its software platform beyond mobile games into enterprise marketing automation, and leveraging its data-rich advertising stack for AI-driven optimization โ both of which could unlock multiple years of EPS expansion
Despite its potential and AXON rollout, $APP is not for the faint-hearted, as its extreme volatility can test even the most resolute shareholdersโ conviction in the business
Additionally, $APP also comes with a high-level of uncertainty
$APP is a good business & appears to be a fair consideration for today at $589๐ต
___
๐๐๐๐๐๐๐๐๐๐โผ๏ธ: ๐๐ก๐ข๐ฌ ๐ข๐ฌ ๐๐๐ ๐๐ง๐ฏ๐๐ฌ๐ญ๐ฆ๐๐ง๐ญ ๐๐๐ฏ๐ข๐๐. ๐๐๐๐ฒ๐ฅ๐จ๐ง ๐๐๐ฉ๐ข๐ญ๐๐ฅยฎ ๐๐ง๐ ๐ข๐ญ๐ฌ ๐ซ๐๐ฉ๐ซ๐๐ฌ๐๐ง๐ญ๐๐ญ๐ข๐ฏ๐๐ฌ ๐ฆ๐๐ฒ ๐ก๐๐ฏ๐ ๐ฉ๐จ๐ฌ๐ข๐ญ๐ข๐จ๐ง๐ฌ ๐ข๐ง ๐ญ๐ก๐ ๐ฌ๐๐๐ฎ๐ซ๐ข๐ญ๐ข๐๐ฌ ๐๐ข๐ฌ๐๐ฎ๐ฌ๐ฌ๐๐ ๐ข๐ง ๐ญ๐ก๐ข๐ฌ ๐ญ๐ฐ๐๐๐ญ.
๐๐ก๐ ๐ข๐ง๐๐จ๐ซ๐ฆ๐๐ญ๐ข๐จ๐ง ๐๐จ๐ง๐ญ๐๐ข๐ง๐๐ ๐ข๐ง ๐ญ๐ก๐ข๐ฌ ๐ญ๐ฐ๐๐๐ญ ๐ข๐ฌ ๐ข๐ง๐ญ๐๐ง๐๐๐ ๐๐จ๐ซ ๐ข๐ง๐๐จ๐ซ๐ฆ๐๐ญ๐ข๐จ๐ง๐๐ฅ ๐ฉ๐ฎ๐ซ๐ฉ๐จ๐ฌ๐๐ฌ ๐จ๐ง๐ฅ๐ฒ ๐๐ง๐ ๐ฌ๐ก๐จ๐ฎ๐ฅ๐ ๐ง๐จ๐ญ ๐๐ ๐๐จ๐ง๐ฌ๐ญ๐ซ๐ฎ๐๐ ๐๐ฌ ๐ข๐ง๐ฏ๐๐ฌ๐ญ๐ฆ๐๐ง๐ญ ๐๐๐ฏ๐ข๐๐ ๐ญ๐จ ๐ฆ๐๐๐ญ ๐ญ๐ก๐ ๐ฌ๐ฉ๐๐๐ข๐๐ข๐ ๐ง๐๐๐๐ฌ ๐จ๐ ๐๐ง๐ฒ ๐ข๐ง๐๐ข๐ฏ๐ข๐๐ฎ๐๐ฅ ๐จ๐ซ ๐ฌ๐ข๐ญ๐ฎ๐๐ญ๐ข๐จ๐ง. ๐๐๐ฌ๐ญ ๐ฉ๐๐ซ๐๐จ๐ซ๐ฆ๐๐ง๐๐ ๐ข๐ฌ ๐ง๐จ ๐ ๐ฎ๐๐ซ๐๐ง๐ญ๐๐ ๐จ๐ ๐๐ฎ๐ญ๐ฎ๐ซ๐ ๐ซ๐๐ฌ๐ฎ๐ฅ๐ญ๐ฌ.
๐๐ง๐๐จ๐ซ๐ฆ๐๐ญ๐ข๐จ๐ง ๐๐จ๐ง๐ญ๐๐ข๐ง๐๐ ๐ข๐ง ๐ญ๐ก๐ข๐ฌ ๐ญ๐ฐ๐๐๐ญ ๐ก๐๐ฌ ๐๐๐๐ง ๐จ๐๐ญ๐๐ข๐ง๐๐ ๐๐ซ๐จ๐ฆ ๐ฌ๐จ๐ฎ๐ซ๐๐๐ฌ ๐๐๐ฅ๐ข๐๐ฏ๐๐ ๐ญ๐จ ๐๐ ๐ซ๐๐ฅ๐ข๐๐๐ฅ๐, ๐๐ฎ๐ญ ๐ข๐ฌ ๐ง๐จ๐ญ ๐ ๐ฎ๐๐ซ๐๐ง๐ญ๐๐๐ ๐๐ฌ ๐ญ๐จ ๐๐จ๐ฆ๐ฉ๐ฅ๐๐ญ๐๐ง๐๐ฌ๐ฌ ๐จ๐ซ ๐๐๐๐ฎ๐ซ๐๐๐ฒ.
tweet
Offshore
Photo
Investing visuals
Iโm close to finishing my first ever deep dive covering $RBRK.
Iโve been working on it for over a month and created more than 20 unique visuals to tell the story of this fascinating business.
Itโll be out this Sunday. Excited to finally share it with you all! https://t.co/JcPGFX65uT
tweet
Iโm close to finishing my first ever deep dive covering $RBRK.
Iโve been working on it for over a month and created more than 20 unique visuals to tell the story of this fascinating business.
Itโll be out this Sunday. Excited to finally share it with you all! https://t.co/JcPGFX65uT
tweet
Offshore
Photo
Investing visuals
Packed earnings week ahead!
Notable names:
โข Monday: $NXP, $WM
โข Tuesday: $SOFI, $UNH, $PYPL, $V
โข Wednesday: $META, $MSFT, $GOOGL, $NOW, $MELI
โข Thursday: $AMZN, $AAPL, $COIN, $NET
Which one are you most looking forward to? https://t.co/jZBZEnHX0g
tweet
Packed earnings week ahead!
Notable names:
โข Monday: $NXP, $WM
โข Tuesday: $SOFI, $UNH, $PYPL, $V
โข Wednesday: $META, $MSFT, $GOOGL, $NOW, $MELI
โข Thursday: $AMZN, $AAPL, $COIN, $NET
Which one are you most looking forward to? https://t.co/jZBZEnHX0g
tweet
Offshore
Photo
App Economy Insights
Tesla earnings breakdown $TSLA
๐ Autonomy's messy timeline
๐ฆพ Optimus 1M product line
๐ The "mirage US rebound
๐ฎ Ambition vs. reality
https://t.co/BowrgD57mt
tweet
Tesla earnings breakdown $TSLA
๐ Autonomy's messy timeline
๐ฆพ Optimus 1M product line
๐ The "mirage US rebound
๐ฎ Ambition vs. reality
https://t.co/BowrgD57mt
tweet
Offshore
Photo
Ahmad
> be us
> Larry & Sergey
> a dorm in Stanford, caffeine shakes, wild ambition
> slap together a janky web crawler
> accidentally organize the entire internet
> call it Google
> build search, email, maps, docs, OS, phones, browser, car, satellite, thermostat, AI lab, TPU farm, and quantum computer
> 2025
> everyone talking about AGI
> OpenAI: โwe need data, sensors, feedback, and scaleโ
> us: staring at Google Maps, YouTube, Gmail, Android, Waymo, Pixel, Fitbit, Docs, Calendar, Street View, and Earth Engine
> "damn. guess we already did that."
> YouTube: 2.6M videos/day
> Android: 3B phones, streaming sensor data 24/7
> Gmail: 1.8B inboxes of human priors
> Search: global-scale RLHF
> Waymo: 71M miles of real-world self-driving footage
> Google Earth: modeled the entire planet
> also your calendar
> people training LLMs on books and PDFs
> we train on humanity
> every click, swipe, tap, misspelled search, scroll, and bookmark
> feedback loop from hell (or heaven)
> depends who you ask
> OpenAI: โwe need $100B for GPUsโ
> us: already built TPUs
> custom silicon
> datacenters pre-co-located with planetary data lakes
> no egress, no latency
> just vibes and FLOPs
> coders: fine-tuning on GitHub repos
> us: 2 BILLION lines of internal code
> labeled, typed, tested
> every commit is a training signal
> Code LLMs dream of being our monorepo
> AGI recipe?
> multimodal perception
> real-world feedback
> giant codebase
> scalable compute
> alignment signals
> embodied sensors
> user data for days
> yeah weโve had that since like 2016
> no investor decks
> no trillion-dollar hype rounds
> just a 25-year accidental simulation of Earth
> running in prod
> OpenAI raises $1T to build AGI
> investors call it revolutionary
> us: quietly mapping 10M new miles in Street View
> syncing another 80PB of Earth imagery
> collecting another year of Fitbit biosignals
> enjoy your foundation model
> we OWN the foundation
> people: โbut Google is fumblingโ
> true
> weโre fumbling in 120 countries simultaneously
> with the greatest compute footprint and research team on Earth
> fumble hard enough and you loop back into winning
> AGI?
> we donโt need to build it
> itโs already inside the building
> powered by Chrome tabs and doc revisions
> mfw we spent 20 years indexing reality
> mfw our data is so good it scares us
> mfw the only thing stopping us from AGI is a meeting between four VPs and one confused lawyer
> call it research
> call it scale
> call it โplanetary simulation-as-a-serviceโ
> we call it Tuesday
tweet
> be us
> Larry & Sergey
> a dorm in Stanford, caffeine shakes, wild ambition
> slap together a janky web crawler
> accidentally organize the entire internet
> call it Google
> build search, email, maps, docs, OS, phones, browser, car, satellite, thermostat, AI lab, TPU farm, and quantum computer
> 2025
> everyone talking about AGI
> OpenAI: โwe need data, sensors, feedback, and scaleโ
> us: staring at Google Maps, YouTube, Gmail, Android, Waymo, Pixel, Fitbit, Docs, Calendar, Street View, and Earth Engine
> "damn. guess we already did that."
> YouTube: 2.6M videos/day
> Android: 3B phones, streaming sensor data 24/7
> Gmail: 1.8B inboxes of human priors
> Search: global-scale RLHF
> Waymo: 71M miles of real-world self-driving footage
> Google Earth: modeled the entire planet
> also your calendar
> people training LLMs on books and PDFs
> we train on humanity
> every click, swipe, tap, misspelled search, scroll, and bookmark
> feedback loop from hell (or heaven)
> depends who you ask
> OpenAI: โwe need $100B for GPUsโ
> us: already built TPUs
> custom silicon
> datacenters pre-co-located with planetary data lakes
> no egress, no latency
> just vibes and FLOPs
> coders: fine-tuning on GitHub repos
> us: 2 BILLION lines of internal code
> labeled, typed, tested
> every commit is a training signal
> Code LLMs dream of being our monorepo
> AGI recipe?
> multimodal perception
> real-world feedback
> giant codebase
> scalable compute
> alignment signals
> embodied sensors
> user data for days
> yeah weโve had that since like 2016
> no investor decks
> no trillion-dollar hype rounds
> just a 25-year accidental simulation of Earth
> running in prod
> OpenAI raises $1T to build AGI
> investors call it revolutionary
> us: quietly mapping 10M new miles in Street View
> syncing another 80PB of Earth imagery
> collecting another year of Fitbit biosignals
> enjoy your foundation model
> we OWN the foundation
> people: โbut Google is fumblingโ
> true
> weโre fumbling in 120 countries simultaneously
> with the greatest compute footprint and research team on Earth
> fumble hard enough and you loop back into winning
> AGI?
> we donโt need to build it
> itโs already inside the building
> powered by Chrome tabs and doc revisions
> mfw we spent 20 years indexing reality
> mfw our data is so good it scares us
> mfw the only thing stopping us from AGI is a meeting between four VPs and one confused lawyer
> call it research
> call it scale
> call it โplanetary simulation-as-a-serviceโ
> we call it Tuesday
tweet
Offshore
Photo
Ahmad
โThe man who Buys a GPU is a fool, until everyone else is stuck with the nerfed models.โ
โ Ancient Llama proverb https://t.co/kHwtmQSyR1
tweet
โThe man who Buys a GPU is a fool, until everyone else is stuck with the nerfed models.โ
โ Ancient Llama proverb https://t.co/kHwtmQSyR1
tweet