findjashua

6 ms·

1.
by findjashua 1mo ago
You're absolutely right!
2.
by findjashua 4mo ago
nasdaq & russell, yes. but not s&p - that has a 6-month requirement for eligibility, which ensures stock is past the lock-up period.
3.
by findjashua 4mo ago
almost all 401k plans offer funds based on s&p 500, not nasdaq/russell others. s&p has also halved their trading days requirement from 1 yr to 6 months, but that's still sufficient to be past the post-ipo lock-up period.
4.
by findjashua 4mo ago
It's not even a fair comparison - the effort involved in learning and building stuff from the internet (in the pre-LLM era) is an order of magnitude higher than turning your OAI/Anthropic account into an absolute slop cannon. I&#x
5.
by findjashua 7mo ago
won't this essentially disable prompt caching, that you get from a standard append-only chat history?
6.
by findjashua 7mo ago
failed the car wash test. i think instead of postiioning as a general purpuse reasoning model, they'd have more success focusing on a specific use case (eg coding agent) and benchmark against the sota open models for the use case (eg q
7.
by findjashua 7mo ago
you need a reviewer agent for every step of the process - review the plan generated by the planner, the update made by the task worker subagent, and a final reviewer once all tasks are done. this does eat up tokens _very_ quickly though :(
8.
by findjashua 9mo ago
providers' ToS explicitly states whether or not any data provided is used for training purposes. the usual that i've seen is that while they retain the right to use the data on free tiers, it's almost never the case for paid
9.
by findjashua 9mo ago
20-yr CAGR seems to be consistently higher than SP500: https://testfol.io/?s=7boYMdNxqjh
10.
by findjashua 10mo ago
i workshop a detailed outline w it first, and once i'm happy w the plan/outline, i let it run while i go do something else
11.
by findjashua 10mo ago
NME at all - 5.1 codex has been the best by far.
12.
by findjashua 11mo ago
i think the standard recommendation is to do range partitioning on the hash of the key, aka hash range partitioning (i know yugabyte supports this out of the box, i'd be surprised if others don't). this prevents the situation of a
13.
by findjashua 1y ago
RAG != EBR
14.
by findjashua 1y ago
why can't you create separate git worktrees, and open each worktree in a separate IDE window? then you get the same functionality, no?
15.
by findjashua 1y ago
the broker should be adjusting the cost basis in the 1099 to account for wash sale. are you not seeing that? what broker are you using?
16.
by findjashua 2y ago
I had checked out SPR, but found the workflow based around reordering history via interactive rebase unintuitive & clunky. That was the motivation behind building this on my own. I have been using this for a few months now, and it has s
17.
by findjashua 2y ago
I miss Phabricator from my time at Meta so much, I made this to achieve a Phabricator-like stacked-commit experience via the git cli: https://pypi.org/project/stacksmith/
18.
by findjashua 2y ago
Shameless self-promotion: https://pypi.org/project/stacksmith/
19.
by findjashua 2y ago
wrong
20.
by findjashua 2y ago
Every single person I know who pays for an LLM is a developer who pays for Claude because of coding ability
21.
by findjashua 2y ago
for a lot of tasks that aren't as cut & dry, i often find myself having to provide it pseudo code, which it can then one-shot to working code. don't get me wrong, it's still a massive upgrade from the pre-sonnet era, but
22.
by findjashua 3y ago
+1 for Sceptre. Mine is coming up on 7 years, and I've never had any maintenance issues w it.
23.
by findjashua 3y ago
hash based partitioning makes repartitioning very expensive. most distributed DB now use key-range based partitioning. Iirc, Dynamo which introduced this concept has also made the switch
24.
by findjashua 3y ago
LM Studio is the easiest way to do it
25.
by findjashua 3y ago
did a comparison on LM Studio - the answers are eerily similar, but Mixtral is way way faster. Codellama-70B is slow to the point of being unusable. (M1 Max, 64 GB RAM)
26.
by findjashua 3y ago
do you find Mixtral to be better than the new 70B one that Meta released a couple days back as well?
27.
by findjashua 3y ago
is this targeted towards non-technical users? for technical users, wouldn't an IDE plugin be a more natural part of the workflow?
28.
by findjashua 3y ago
phabricator & scuba are far better than anything i've used anywhere else.
29.
by findjashua 3y ago
i mean it’s Andreesen. They’re like the Softbank of Silicon Valley
30.
by findjashua 4y ago
pyright

More ›