amasad

7 ms·

1.
Replit (YC W18) Is Hiring(jobs.ashbyhq.com)
1 points by amasad 8mo ago
2.
by amasad 9mo ago
We do let you change it; under your account setting there is a email with an edit button next to it. And it shouldn’t need to match your GitHub email. Perhaps there is some weirdness if you’ve signed up with GitHub. Feel free to email me an
3.
Replit (YC W18) Is Hiring Engineers to Automate Coding(replit.com)
1 points by amasad 1y ago
4.
Using Tvix Store to Reduce Nix Storage Costs by 90%(blog.replit.com)
2 points by amasad 2y ago|0 comments
5.
Replit (YC W18) Is Hiring Software Engineers in the Bay Area(replit.com)
1 points by amasad 2y ago
6.
Skip the README, let us install for you(blog.replit.com)
1 points by amasad 3y ago|0 comments
7.
Prioritizing Employee Liquidity at Replit(blog.replit.com)
1 points by amasad 3y ago|0 comments
8.
by amasad 3y ago
We're on a budget :) trained on 128 H100-80GB GPUs for a week (200B tokens over 5 epochs, ie 1T tokens). Tech talk here with timestamp: https://www.youtube.com/live/veShHxQYPzo?si=UlcU9j2kC-C4oWvj...
9.
by amasad 3y ago
Crates is especially challenging -- the build process is very expensive on all resources (cpu, ram, and disk), and packages are very hard (impossible?) to cache. It works super well on the Pro plan.
10.
by amasad 3y ago
It’s communication meant to existing users and it’s a nuanced picture worth elaborating. But if you are not an active user it’s best if you just look at our pricing page for a complete picture: https://replit.com/pricing
11.
by amasad 3y ago
This is for hosting. Development remains free
12.
by amasad 3y ago
Yes, in fact we recently made the free plan more powerful with non-preemptible VMs, premium networking, and package caching for everyone. And working on an AI free tier coming soon. Will likely invest more in the free plan once the abuse fr
13.
AI Agent Code Execution(blog.replit.com)
1 points by amasad 3y ago|0 comments
14.
Replit ModelFarm: Zero-Config LLM App Dev Environment(blog.replit.com)
2 points by amasad 3y ago|0 comments
15.
Autoscale and Static Deployments(blog.replit.com)
1 points by amasad 3y ago|0 comments
16.
Speeding Up Deployments with Lazy Image Streaming(blog.replit.com)
5 points by amasad 3y ago|0 comments
17.
Replit (YC W18) Is Hiring Engineers in AI, Platform, and Product(replit.com)
1 points by amasad 3y ago
18.
by amasad 3y ago
the blog is slightly outdated — 3000 is out of beta and can be used now. Here is a fun game built with it: https://replit.com/@slmjkdbtl/Bean-Survivor
19.
Kaboom.js 3000(kaboomjs.com)
140 points by amasad 3y ago|13 comments
20.
Local Code Chatbot Running on 2GB RAM(twitter.com)
1 points by amasad 3y ago|1 comments
21.
Replit (YC W18) Is Hiring Engineers and Designers(replit.com)
1 points by amasad 3y ago
22.
by amasad 3y ago
Is this the best you can find? not even top 10 bangers.
23.
by amasad 3y ago
Interesting. This seems like a weakness of natural language understanding. If you rephrase your prompt slightly it would get it right. Try: // return even numbers that are also more than 10 const arrayFilter = (array) =>
24.
by amasad 3y ago
Yes we have a robust extension system and some are already building alternatives.
25.
by amasad 3y ago
Broadly finetuning is any post pretraining training. Most of the time it is used in the context of fitting a more narrow task. In our case, it was the same training objective as the pretraining but meant to be more representative of what Re
26.
by amasad 3y ago
The model is not RLHF'd or instructed. It's an inline autocomplete model so it will get confused if you talk it like you're talking to a person. Altho it is possible to finetune it this way. To get better full function comple
27.
by amasad 3y ago
Can't find it now but pretty sure BigCode said somewhere they explicitly looked for it and removed it. Also subjective measure does match up to the benchmark. Our finetuned model performed +50% on HumanEval and then when using it felt
28.
by amasad 3y ago
1- We trained on languages that are most popular on Replit. Markdown is important because you need some amount of natural language in the data, and it will act as a sort of "natural language label" for code. 2- I like how portable
29.
by amasad 3y ago
Hopefully on phones too
30.
by amasad 3y ago
it is 2.7B

More ›