wrhn

hashingroll

Joined 6 years ago

Karma
129
Submissions
38
Member since
Jul 10
I graduated in 2020 with a CS degree and work as Software Engineer at Google ever since. I like coding, programming languages, algorithms, and solving puzzles. At Google I have worked on AI/ML to solve Spelling & Grammar and Tone & Style suggestions in Google Docs and Gmail. Right now I work on modeling LLMs to bring Generative AI features in Google Docs.
View on Hacker News

Recent submissions

Comment · 4 years ago
This might a long read but it is well-written reader-friendly analysis of Tarjan's proof with chain of reasoning. https://codeforces.com/blog/entry/98275
Comment · 4 years ago
Not all optimizations are more energy consuming. For an analogy, does a using a car consume more energy than a bicycle? Yes. But a using a bicycle does not consume more energy than a man running on feet.
Comment · 4 years ago
Love this! Took a while to find the solution :)
Comment · 4 years ago
In the same vein, an auto-initiater would work like

Me: <Types question and hit send>

(Doesn't actually get sent yet)

My auto-initiater: Hi

Other person: Hi, <...>

(no chat notification, the original question gets sent now)

Other person: <Answers/Whatever>

(gets chat notification)

Comment · 4 years ago
I would not discount NLP tasks just yet. In practice, they are still used to solve problems like spellcheck, autocomplete and even words and text rewrites.
Comment · 4 years ago
The state of the art models these days would not be able to be bundled with browsers. These are typically neural models [0] and require specialized hardware to run, and also have high memory requirements. The models in browsers are light-weight to be able to run on most devices. Might be similar to those used in smartphone keyboards like GBoard [1].

[0] https://www.microsoft.com/en-us/research/blog/speller100-zer..., https://cloud.google.com/blog/products/g-suite/everyday-ai-b...

[1] https://ai.googleblog.com/2021/10/grammar-correction-as-you-...

Comment · 4 years ago
Comparing the spellcheck quality of Firefox to Google (search, not Chrome) might not be completely fair. The browser most likely uses a client-side spelling model with tight memory & performance constraints. The suggestions from Google come from high quality models hosted on their servers.
Comment · 4 years ago
Not sure about TV shows but for films (especially Netflix originals), that's partly due to the recent trend of films getting shot for TV viewing experience. These films tend to have fewer "wide" shots and more "character" shots compared to the earlier theatre release only films.
Comment · 5 years ago
> no mutable reference parameters

Sidenote: The style guide (recently?) removed the ban on non-const references [0]. They are now allowed for non-optional output parameters. Though returning value is generally preferred over output parameters.

[0] https://google.github.io/styleguide/cppguide.html#Inputs_and...

Comment · 5 years ago
The argument for tabs really goes that time is spent much more on reading code than authoring, so it should be optimized and customizable (tab width) for the reader.
Comment · 5 years ago
> Convert comments to code. Write a comment describing the logic you want, and let GitHub Copilot assemble the code for you.

I don't know if Copilot does it already but I would love if there was a tool that does exactly the opposite — convert code to comments and documentation.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton