1/ A few select excerpts from Lux Q2 letter to LPs…
“The best hedge against running out of cash—is raising it”
2/
-“Question your teaspoon”
-“While many talk of discouraging ‘conflicts of interest’
few appreciate and encourage ‘interest in conflicts’
-The Binary (dichotomous, stark contrasts)
-The Blur (dialectic, smooth continnum)
3/ “We don’t mistake for genius what is partly attributable to central banks buying …driving rates low…and raising asset prices”
“Means may move (upward) but a reversion to the mean (downward) may move away people’s means….
4/ “Eventually the proverbial bath water will turn from TAINTED to TOXIC
& babies may not just be thrown out with the bathwater––but forcibly hurled”
The effect will eventually be the cause of low returns
We know that eventually “eventually” will arrive—but not when “when” will
5/ Lux is increasingly backing entrepreneurs hailing from—or building—overseas…
London, Dublinm Berlin…
The sun hits first and the day starts ahead in Europe—as do many trends from hairstlyes (via soccer stars) to regulatory watchdog decrees (from GDPR to new AI regs)…
6/ Decentralizing + Recentralizing
-Mechanical fans moved off our computers and en masse into server farms
indentured servitude of machines of loving grace
on China v US
-Dependence demands concessions + compromises
-Independence supplies indifference + comprises self-reliance
• • •
Missing some Tweet in this thread? You can try to
force a refresh
1/ On Feb 21 I told my boss aka my wife @LTwolfe that I had my third conviction call not yet appreciated + to buy a basket
*In 2016 it was NVDA, pitched publicly at Invest For Kids in Chicago (80x+) when Lux portfolio company Zoox was using NVDA chips and I predicted the narrative would change from gaming consoles PS4/Xbox to simulation and AI…
3/ In the fall is 2024 I pitched SK Hynix, Micron, Samsung and we bought in our pa
Mostly because I had read a technical paper about LLMs and inference on device with local flash (again changing narrative from commodity memory players to coveted attach rates on HBM AND edge)
1/ New paper from @ylecun et al on alternative approach for AI to learn more biologically...
paper basically says AI is super smart but still can't learn like a toddler can...
the main critique
the main critique...
is that top models today dont learn once deployed.
and curating data + training pipelines + filtering out noise is all done by humans in loop instead of being an intrinsic skill of the AIs...
here's Claude explaining the paper simply
3/ not quite Kahneman style but they call it System A +B
System A = learning from observation
this is self-supervised learning like predicting next word or video frame. good at making representations but passive and struggles to understand cause + effect...