HackerTrans
TopNewTrendsCommentsPastAskShowJobs

arunaugustine

no profile record

Submissions

Notes on the Math of LLMs (PDF File Link)

drive.google.com
1 points·by arunaugustine·11 เดือนที่ผ่านมา·0 comments

comments

arunaugustine
·7 เดือนที่ผ่านมา·discuss
They had a scheduled maintenance between 7am and 11am UTC in Chicago. But that should have re-routed traffic not take down internet right?
arunaugustine
·7 เดือนที่ผ่านมา·discuss
Shopify is down.
arunaugustine
·9 เดือนที่ผ่านมา·discuss
For someone who wants to tread this path, using postgreSQL stored procedures in a team settings, what would be a good dev workflow in a team. Update them in git and use a CI to update the DB etc? Is there some tips that you can share on that?
arunaugustine
·ปีที่แล้ว·discuss
This is exactly what I loved about the Causal app (no affiliation). They started as a general purpose spreadsheet with 'Amb' cells built-in, though later on they seem to have converged on the financial modeling space.

[0]: https://causal.app/
arunaugustine
·2 ปีที่แล้ว·discuss
On the bright side, you need to pay only $1+, if you are using it commercially. Seems a reasonable ask and a nice way to support the development. Interesting approach to licensing indeed, but I wouldn't call it messed up. It's just new.
arunaugustine
·2 ปีที่แล้ว·discuss
TinyBase looks very promising! Is there a doc or reference you can guide me to, for using TinyBase with Preact.js instead of React?
arunaugustine
·2 ปีที่แล้ว·discuss
Probably referring to Zig migrating away from AWS to become more cost efficient [1]. This was also mentioned in the announcement regarding this donation from the Zig Software Foundation [2]

[1]: https://ziglang.org/news/migrate-to-self-hosting/ [2]: https://ziglang.org/news/300k-from-mitchellh/
arunaugustine
·2 ปีที่แล้ว·discuss
CUE lang [1] is one instance where the idea of data types as lattices is implemented in industry

[1]: https://cuelang.org/docs/concept/the-logic-of-cue/#the-value...
arunaugustine
·2 ปีที่แล้ว·discuss
Also an Ex-FANG software engineer.

1. Quit and returned to home country before starting up. I did try the evenings and weekend startup approach, but got close to burn out. For me at least, starting up meant working on something full time.

2. Took 4 pivots and as many years to get our first dollar. In year 2, we did earn some money but was cost of goods sold (travel tickets, building materials, do can't really count that as dollar earned).

3. It was a series of plateaus. Took a long time for the first $10, then about a year to first $1000, then an year to $3000 and then another year to $10K. This is just one data point. Smarter/Luckier people may do it faster/better.

4. I love software engineering. I am still doing it at my startup. I'm doing it my way, learning and experimenting with stuff. Actually my core motivation is to keep doing software (and someday even build a programming language) supported by a team and company creating value for others and getting paid for it.

5. With a co-founder.

6. An article on the initial journey, I wrote down a while back. https://medium.com/messengermarketing/from-quitting-silicon-...

Btw, the idea mentioned at the end of the article also failed. But that final pivot is what we are still working on (in-cart campaign software for Shopify stores). This second half of the journey is still ongoing. Will try to write it down someday.