valuedbuffett

Building @valuedterminal

[valuedbuffett @valuedbuffett]

In a gold rush, you don’t want to chase the gold; you want to sell the shovels.

[valuedbuffett @valuedbuffett]

@valatility @paradex nice hedge

[valuedbuffett @valuedbuffett]

bullish on china 🇨🇳

[valuedbuffett @valuedbuffett]

@valatility Parissssss

[Lit Nomad @LitNomadT]

At my hedge fund, a quant introduced a silent bug into our trading system that cost us $5M (in forgone profits) over a few months before anyone noticed. The automated trading system just seemed a little less aggressive than usual, so it took a while before our head trader asked us to audit the system, which uncovered the issue.

The quant who introduced the bug was well-respected and brilliant, so he knew he wouldn’t get fired. But he left for a competitor a few months later because he knew they’d take a big chunk out of his bonus that year. This perverse incentive structure always bothered me.

Before becoming a quant, I was a physics student, so I was in the habit of solving long and complex problems that filled up pages with equations. If you make one mistake early, the error compounds from there, and all the downstream work is useless. Because of this, I was always very careful when committing code changes to live quant models and made sure to understand everything from first principles. I’d be scolded for being “too careful” by teammates who just wanted to “fail forward” to get a big bonus.

The hidden issue with this mindset is that we are often drawing false conclusions based on a faulty foundation and never find out. Humans rarely get 100% correct on a tough physics exam, so we know there are likely bugs in our research findings if we rush through the work. But unlike academia, there isn’t an answer key to grade the exam. The bugs can go unnoticed forever, and often do. Entire models are retired when they could have been wildly profitable if someone had noticed the bug. The true cost of these bugs is immeasurable.

These perverse incentives exist across many domains outside of quant finance and create unnatural ceilings on progress and research. In graduate school, when I worked on image processing research for tomographies, I was also shocked by how many human errors went unnoticed. But I just wanted to get my paper published and move on. Having worked deep in the bowels of different codebases (defense industry, academia, quant finance), you’d be shocked by how many human errors exist across all these domains.

One of the greatest benefits of AI is that it will audit entire systems and fix these silent bugs, removing these unnatural ceilings and unlocking boundless potential. I think R&D progress across the sciences and engineering is currently underpriced and is going to explode exponentially. We’re going to see that play out in the S&P 500 over the next decade.

[valuedbuffett @valuedbuffett]

@lightningkoray kuss diggi

[valuedbuffett @valuedbuffett]

@remarks this is so funny hahahhah

[valuedterminal @valuedterminal]

One terminal.

The edge is on one screen now.

[valuedbuffett @valuedbuffett]

holy overreaction

[valuedterminal @valuedterminal]

Two of thirty-one October sessions produced Bitcoin's entire 7.34% loss. The other twenty-nine finished flat. A quiet month can still lose 7%. Here is what it actually was, measured rather than narrated. General information only, not investment advice. https://t.co/KoJJvmi0lY

[valuedbuffett @valuedbuffett]

alpha for crypto traders

[valuedterminal @valuedterminal]

One terminal.

The edge is on one screen now.

[SalsaTekila @SalsaTekila]

This is bullish BTC, probably bearish HYPE

[valuedbuffett @valuedbuffett]

$SPY

Bottom in.

[valuedbuffett @valuedbuffett]

he is just playing with the market…

[valuedbuffett @valuedbuffett]

@q_79t

[valuedbuffett @valuedbuffett]

That is why you should buy…

Not Financial Advice

[valuedbuffett @valuedbuffett]

$MSTR $STRC

@Strategy is doing a great job.
That is also why analysts are
so bullish on that stock.

LFG

Not Financial Advice

[valuedbuffett @valuedbuffett]

@ShaneStoffer Goodbye 👋