ashu1461

9 ms·

121.
by ashu1461 1y ago
Landing page is indeed very refreshing
122.
by ashu1461 1y ago
A huge supporter of typing in python, but have observed that type sense which pycharm provides out of the box has its fair share of bugs as well. Also should check out ty by astral, it is pretty fast and does a good job at typechecking. ht
123.
by ashu1461 1y ago
This Without typing it is literally 100x harder to refactor your code, types are like a contract which if are maintained after the refactor gives you confidence. Over time it leads to faster development
124.
by ashu1461 1y ago
Sometimes that is not an option, consider the bias which python gets due to all the LLM libraries being python first.
125.
by ashu1461 1y ago
This surpasses the previous record of 4 months https://www.science.org/content/article/longest-human-transp... Pretty amazing tech tbh
126.
by ashu1461 1y ago
Somehow after looking at the video, it is hard to believe that it is actually much different than your standard knife.
127.
by ashu1461 1y ago
Is it weird I went through the complete landing page and still did not get what actually the features are
128.
by ashu1461 1y ago
The assumption that every programming round is a b tree reversal is not right, and I think programming rounds tell a lot about the candidates thought process, how they structure the code, how they handle edge cases and a lot of other things
129.
by ashu1461 1y ago
This definitely helps, but what about the programming or implementation rounds ?
130.
Ask HN: What are good questions to ask in a remote round in post GPT era?
7 points by ashu1461 1y ago|18 comments
131.
by ashu1461 1y ago
While the idea is good, can someone share the list of use cases which can be built using this please ?
132.
by ashu1461 1y ago
What makes you think it’s hard? We have AI generating songs and writing code, but setting up basic health checks is too much?
133.
by ashu1461 1y ago
How do we know that the program is failing ? How hard is it for the frontend to detect if the last update to the status page was made a while ago and that itself implies there is an error and should be reported ?
134.
by ashu1461 1y ago
GPT is working in agent mode, which kind of confirms that claude is hosted on google and GPT probably on MSFT servers / self hosted.
135.
by ashu1461 1y ago
So even then, it should have been able to correctly report the status, it somehow shows that the status page is not automated and any change there needs to go through someone manual.
136.
by ashu1461 1y ago
Just proves how shady the status page and sla stuff is
137.
by ashu1461 1y ago
I think multiple regions are down. asia-south, us-east atleast are impacted.
138.
by ashu1461 1y ago
When you deploy code generated by Gemini :D
139.
by ashu1461 1y ago
I don't think so, it might be true for their long tail customers, but most of the tech folks have not used Google Search / Gemini APIs in ages.
140.
by ashu1461 1y ago
Have to work hard to figure out the difference between - Vertex AI - AI Studio - Gemini - Firebase Gen AI
141.
by ashu1461 1y ago
There is also no way to over-write content moderation settings, and half of the responses you generate via open ai endpoint end up being moderated.
142.
by ashu1461 1y ago
How about calling 1-800-CHATGPT (1-800-242-8478) now ? Wonder how many queries which the university is calling can now be automated
143.
by ashu1461 1y ago
Tesla Model S is also not exempted ?
144.
by ashu1461 1y ago
Is it right to assume that the PATH env variable and the context in what it is used are two different things ? While the PATH variable fundamentally is same as other env variables like HOME / USER but how PATH is interpreted will chang
145.
by ashu1461 1y ago
Would a LLM under human guidance turn out to be a good verifier ? i.e. if LLM knows the rules to verify or has enough data points (internet access, actual responses)
146.
by ashu1461 1y ago
Have one question So if snapshot violation is happening inside Multi-AZ instances, it can happen with a single region - multiple read replica kind of setup as well ? But it might be easily observable in Multi-AZ setups because the lag is hi
147.
by ashu1461 2y ago
There have been similar such statements in the past but nothing happened, What will be different this time ?
148.
by ashu1461 2y ago
Are these attached directly to your server or hosted separately ?
149.
by ashu1461 2y ago
One place where I feel gemini models lag is function calling and predicting correct arguments to function calls, is there a benchmark which scores models on the basis of this ?
150.
by ashu1461 2y ago
They mention an user taking an action will be billed. I guess even sending a message or reacting with an emoji would count as taking an action ? Even logging in ?

More ›