Hacker News: carlossouza

New comment by carlossouza in "Amazon CEO's talks with U.S. officials triggered crackdown on Anthropic models"

carlossouza — Sun, 14 Jun 2026 02:03:59 +0000

I'm just stating facts:

- Amazon's CEO knew what he was doing and the possible consequences

- Anthropic must raise cash, and there are only 2 ways: an IPO or follow-ons

- If the IPO is blocked, existing investors will be able to increase their stake on Anthropic at a very attractive lower valuation

- Amazon has 20% of Anthropic: so, they benefit from it

New comment by carlossouza in "Amazon CEO's talks with U.S. officials triggered crackdown on Anthropic models"

carlossouza — Sun, 14 Jun 2026 00:08:25 +0000

Of course, Amazon wanted this to happen.

They own 20% of Anthropic.

Anthropic bleeds cash. They have to raise capital.

There are only 2 ways: an IPO or follow-ons from existing investors.

If the IPO gets delayed because of these restrictions, Anthropic will be forced to raise more capital from existing investors.

And existing investors (Amazon) will end up owning more of Anthropic at a cheaper valuation.

New comment by carlossouza in "Statement on US government directive to suspend access to Fable 5 and Mythos 5"

carlossouza — Sat, 13 Jun 2026 14:01:28 +0000

Good for Anthropic?

Can you think of any case in history where the US government suspended product sales due to national security concerns and that was ultimately beneficial for the company being regulated?

New comment by carlossouza in "NSA is using Anthropic's Mythos despite blacklist"

carlossouza — Mon, 20 Apr 2026 13:52:34 +0000

“Show me the incentives and I will show you the outcomes.” Charlie Munger

New comment by carlossouza in "Code has always been the easy part"

carlossouza — Wed, 25 Feb 2026 03:52:33 +0000

What's easy for some might be truly hard for others.

The Bitter Lesson: Applying Deep Learning to Enhance Momentum Trading Strategies

carlossouza — Wed, 23 Apr 2025 12:19:40 +0000

Article URL: https://www.quantitativo.com/p/the-bitter-lesson

Comments URL: https://news.ycombinator.com/item?id=43771245

Points: 1

# Comments: 0

LambdaMART LTR algorithm applied to cross-sectional momentum strategies

carlossouza — Sat, 01 Mar 2025 19:14:14 +0000

Article URL: https://www.quantitativo.com/p/learning-to-rank

Comments URL: https://news.ycombinator.com/item?id=43222591

Points: 1

# Comments: 0

Coding Trend Factor: Implementing a High-Performing Factor from Research

carlossouza — Mon, 10 Feb 2025 16:04:32 +0000

Article URL: https://www.quantitativo.com/p/coding-trend-factor

Comments URL: https://news.ycombinator.com/item?id=43001798

Points: 2

# Comments: 0

Can a good execution algo enable a 28% annual return with a 19% max drawdown?

carlossouza — Sun, 20 Oct 2024 22:20:36 +0000

Article URL: https://www.quantitativo.com/p/mind-the-gap

Comments URL: https://news.ycombinator.com/item?id=41898964

Points: 1

# Comments: 0

Varadi-Oscillator-based mean-reversion strategy with 21% CAR since 2004

carlossouza — Sun, 29 Sep 2024 16:55:10 +0000

Article URL: https://www.quantitativo.com/p/a-different-indicator

Comments URL: https://news.ycombinator.com/item?id=41688616

Points: 2

# Comments: 0

Implement algorithms that minimize slippage

carlossouza — Sun, 15 Sep 2024 16:45:20 +0000

Article URL: https://www.quantitativo.com/p/the-devil-is-in-the-details

Comments URL: https://news.ycombinator.com/item?id=41548547

Points: 62

# Comments: 8

Coding Live Forward Tests

carlossouza — Wed, 04 Sep 2024 11:30:48 +0000

Article URL: https://www.quantitativo.com/p/coding-live-forward-tests

Comments URL: https://news.ycombinator.com/item?id=41444406

Points: 1

# Comments: 0

Long and Short Mean Reversion Machine Learning

carlossouza — Mon, 26 Aug 2024 19:30:05 +0000

Article URL: https://www.quantitativo.com/p/long-and-short-mean-reversion-machine

Comments URL: https://news.ycombinator.com/item?id=41360873

Points: 1

# Comments: 0

Machine Learning and the Probability of Bouncing Back

carlossouza — Wed, 21 Aug 2024 14:39:27 +0000

Article URL: https://www.quantitativo.com/p/machine-learning-and-the-probability

Comments URL: https://news.ycombinator.com/item?id=41310759

Points: 1

# Comments: 0

Trading ETFs while fear and greed rise

carlossouza — Tue, 06 Aug 2024 09:04:00 +0000

Article URL: https://www.quantitativo.com/p/trading-etfs-while-fear-and-greed

Comments URL: https://news.ycombinator.com/item?id=41169018

Points: 2

# Comments: 0

A portfolio of mean-reversion strategies that delivers 26% CAR since 2010

carlossouza — Sat, 27 Jul 2024 19:39:40 +0000

Article URL: https://www.quantitativo.com/p/trading-the-mean-reversion-curve

Comments URL: https://news.ycombinator.com/item?id=41088817

Points: 3

# Comments: 0

Momentum-Based Long and Short Equities Portfolio

carlossouza — Sun, 07 Jul 2024 21:10:51 +0000

Article URL: https://www.quantitativo.com/p/momentum-based-long-and-short-equities

Comments URL: https://news.ycombinator.com/item?id=40900457

Points: 1

# Comments: 0

A mean-reversion strategy with 26% annual returns (multiple instruments)

carlossouza — Sat, 29 Jun 2024 20:17:46 +0000

Article URL: https://www.quantitativo.com/p/robustness-of-the-211-sharpe-mean

Comments URL: https://news.ycombinator.com/item?id=40833094

Points: 7

# Comments: 1

Larry Connors' Cumulative RSI strategy backtest: 26% annual return in 24 years

carlossouza — Mon, 24 Jun 2024 00:58:48 +0000

Article URL: https://www.quantitativo.com/p/squeezing-more-profits-with-cumulative

Comments URL: https://news.ycombinator.com/item?id=40771908

Points: 2

# Comments: 0

New comment by carlossouza in "Changing a mean reversion strategy to deliver 30% annual returns since 1999"

carlossouza — Mon, 17 Jun 2024 12:44:17 +0000

The data used is Sharadar Core US Equities Bundle:

https://data.nasdaq.com/databases/SFA

It's a great survivorship-bias-free dataset.

Regarding tools, I use Python. I wrote the backtesting software many, many, many years ago during my Master's degree, and I've been refining it ever since.

It's an event-driven engine (they are slower than vector-based engines, but they are easier to write strategies for, understand, and debug) with all the bells and whistles, similar to the late Zipline. In fact, I tried most of the Python backtest engines that exist, and that's why I prefer to use what I built over the years: I have 100% understanding of what’s happening and 100% control.

I’m thinking about open-sourcing it… anyway, the logic is not that complicated.