90% of projects fail; even with smart founders and people. Why?
Too often we lack humility, rely on our own judgment and ignore reality.
Michael Mauboussin calls this the “inside view.”
The fix? Use the outside view.
👇 Read more: https://t.co/lgX5kP6n2i
I mostly use codex-high 90% of the time and only xhigh for planning. Just because xhigh is a little too slow for my liking.
Anyone else could convince me why I should use xhigh all the time?
This 1000% updating your dependencies for the sake of updating makes no sense.
Now with AI writing code there's no need to import a 'pluralizeWord' library
Fork your dependencies, trim them to only your use case, never update unless it breaks for your users. I’ve been vocal about this for 10+ years. I’ve always said that updating is way riskier than latent bugs (which can be tracked and CVEs monitored).
If you are updating a dependency, it’s on you to analyze every single commit in the full transitive set of dependencies. If you dont see anything compelling, dont update!
I remember at HashiCorp once in awhile an engineer would try to update a dep or replace a DIY lib with an external one and id always ask “show me the commit we need.” Dont update for the sake of it.
Feeling pretty swell about this mentality with all the supply chain attacks happening.
I'm gonna be honest with you. after work, I used to play games a lot.
these days I feel like I can't even afford to start a new game or entertain myself anymore or watching tv because I'm missing out on tokens or yngmi
now its either doomscrolling or AI go bbrrrrrr...
AI psychosis is real.
Bearish on the entertainment industry
Just shipped Projection Lab to Marketpal.
Every stock page now has a clean 5-year model.
Bear / base / bull cases.
Editable revenue growth, margins, dilution, exit PE & P/S.
Instant price range + implied CAGR.
LULU example below.
Interesting my main takeaways:
- HIIT and fasting before sleep is godsend
- no doom scrolling before sleep
- AC is good
- no more alcohol
- sleep divorce makes no sense if initiated by the wife
Did a few weekly 2 year backtests on my watchlist universe this morning ~ about 95 stocks.
Interestingly enough, pure weekly momentum outperformed the versions that added QVM scoring: quality, value, and momentum. (simple_score_v1, quality_swing_rebuild_v1)
My guess is my watchlist was already quality-filtered by me.
For a stock to make the list, I probably already liked the business, the theme, the setup, or the long-term story.
So adding more quality and value filters didn’t help much.
The edge seem to be rewarding trend-following
perhaps not because quality or value don’t matter
But because inside a curated watchlist, momentum became the better sorting mechanism. I've been mostly a mean reversion guy but this prompted me to experiment more with momentum based strats instead of chasing value.
Step away from your laptop. Keep building with Codex on your phone.
Codex keeps working on your computer, with your files and project context still in place.
Pocket-sized access. Full Codex working state.
https://t.co/vGsw5wnHnG
@theo I like vertical splits view so I can read multiple chats, I got hotkeys to go to diff tabs and I can easily switch to nvim or git diff review tools of my choice when I wanna read code.