I understand prompt caching is an effective way for optimising for input tokens. What are other options for optimising input tokens? My input tokens maximise in using an xml file.I have avoided using TOONs [1] because it works only with uniform arrays, not nested objects nor non-uniform structures.[1]https://github.com/toon-format/toon
The good thing is you are exploring a lot. And exploration never ends until you begin to enjoy getting tired.
I have been using Slack for 7+ years. I have recently joined a startup which communicates on emails and calls. Since we're an early stage, I am considering alternatives to Slack. I love Slack for its deep search. But it could get pricey for our team in the next six months.I am considering Rock (www.rock.so). How has your experience been with integrations and search?
There are some well known api docs such as that of Stripe, Plaid, Twilio, etc. I believe there are some open source API docs too which can be used for inspiration.
An entrepreneur is not a label limited to founders only. I believe every person working on an innovative product is an entrepreneur, irrespective of stage of the company. What are the traits of such people?
e.g. if you repay a 10 year loan in 4 years. You will be rewarded with cash as a percentage of interest that were yet to be paid in next 6 years.
Lenders make money from the good repaying behaviour of good borrowers. However, lenders are known to restructure loans with bad borrowers to recover pending amount. This is a moral hazard signalling the good ones that they aren’t doing any good by repaying on time. Good repayment behaviour has helped lenders, what about the borrowers?
The mathematics of lottery is based on 6/49 game. Each player selects 6 numbers without replacement from an urn of 49. If the 6 numbers match with the lottery number then it is a jackpot. The probability of it happening is 1 in C(49,6) or 1 in 13 million. I have two questions: [1] Why does the urn consists of 49 numbers and not 48 or 50 numbers? [2] Why one has to select 6 numbers and not 5 or...
These questions are about candidates past, however I intend to know the current status of mind. This reminds me that I was able to confirm a candidate's state by asking him about his daily routine. His after-work-hours were very depressing for me. Conversations are serendipitous - how they have an innate quality of bringing out the best from within. :-)
- Downsides of acquihire is that we may not adopt to the culture to build markets. I want to work and build on their product (i am a user). They want to expand to a market that we are targeting and our domain knowledge attracts their team.
- Currently, I really want a direction: our TG is affected by the slowdown. It is also the TG that will emerge first as soon as the markets open. I am unsure how long the delay is going to last. In next few years, if the acquihire turns out to be right then I'll be in the centre of building markets that I love. In the long run, I want to build a useful venture. It will be the last place where team mates come together to build their careers.
- I never thought of actively looking for buyers because I would not buy my own company today. Although we have customers, I think we have a lot to do before we begin to surprise ourselves.
- Two roles that we could hire: This is an interesting point. I don't have to think of those two roles, it is on my mind. As and when we get a hint of getting a bandwidth for those two roles, we'll roll out the JDs.
This is the third acquihire offer we've received in last 3 months and fourth overall. All the previous ones were easy to decline. But this one is making me think for a number of reasons: (a) It's targeting the same user base. And they already have the data that we need to build & scale our existing products. (b) It has got thousands of customers and the product is a monopoly. (c) Backed by...