WasmGPT
AI
local llm chat
SHOW HN2023
185points
50comments
#310of 3,109 in 2023
Top 10%of 2023 by points
How HN took it
RoastedHalf or more of the comments are critical
3 positive, 8 neutral or asking, 2 mixed, 17 critical, out of 30 comments about the project. Thread score -0.42.
Performance: 1 for, 17 againstExecution: 1 for, 0 againstIdea: 1 for, 0 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
Fun to try and impressive that it runs locally in the browser
From 2 comments
Impressive! Just tried and woks fine (linux,i5,chrome)
throwaway888abc on HNregardless of gibberish or not it's pretty amazing that you can put it in the browser as Wasm, locally inferred
jijji on HN
Objections
Output quality is gibberish or nonsensical for most questions
From 3 comments
I tested this using about 10 different questions that were answered correctly on chatGPT, all of the answers produced gibberish... is it because you're using the GPT-2 model?
jijji on HNIt failed to accomplish all the trivial tasks I've requested, rendering it unusable for the time being. ``` User: implement a python function that takes two numbers and return…
hathym on HNProbably having too much fun with this, thanks ``` Human: What is 3+3 Assistant: 3+3 Human: What does it equal Assistant: It is equal to the following: Assistant: It is equal to…
nwbort on HN
Much slower than WebGPU alternatives
From 1 comment
It is not. It's much slower. But it doesn't require you to use a canary browser with a command line flag, so that's that. EDIT: I guess I should give credit here...…
lxe on HN
Requests
Progress indicator while the answer generates
From 2 comments
Nice stuff! A bit of feedback, if I may: * maybe it worth showing some feedback when AI is thinking/composing the answer vs. when it's already finished. Currently, on my machine…
Lex-2008 on HNA progress indicator on the download might’ve been nice
pmarreck on HN
Fix non-Latin and UTF-8 input handling
From 1 comment
Nice stuff! A bit of feedback, if I may: * maybe it worth showing some feedback when AI is thinking/composing the answer vs. when it's already finished. Currently, on my machine…
Lex-2008 on HN
Questions
Is it faster than WebGPU and why
From 1 comment
Is it faster than webgpu? If so, what makes it so?
eachro on HN
Alternatives named
ChatGPT
From 2 comments
:D, good and funny to read indicator of the models "intelligence". It's at least more fun than ChatGPT (I am feeling good and so have I!): User: Hi wasmGPT, how are you?…
teekert on HNSorry for a layman like me, what does this differ with ChatGPT web interface on openAI website? I can't get it what it does.
terrycody on HN