Hacker News Reader

mesmertechon Content Injection Attack on GitHub

that was the first iteration of CSS injection that was working, that github then patched. The new one is the new iteration that still works

it was found by a bunch of anime-pfps on twitter and went "viral"

mesmertechon Europe probes Microsoft's €15M stake in AI upstart Mistral

Love it as a consumer because Mistral is probably not doing any more open weights if they get tangled with Microsoft.

But man, as a business owner in EU, this would just make me not want to build an AI startup in EU

mesmertechon Stable Code 3B: Coding on the Edge

Wondering the same here

mesmertechon OpenAI releases Whisper v3, new generation open source ASR model

Wow a fellow slovak indie developer, kinda rare to see.

mesmertechon Ask HN: Show me your half baked project

https://aieasypic.com/

Have plans for incorporating pretty much all AI image gen. capabilities, but only has image gen, loras, controlnet, background removal and upscale atm. There's like something new in AI image gen coming up everyday and I still haven't gotten lora training, checkpoint training, image colorizing, AnimateDiff and so much more done yet.

mesmertechon New AI experiences across our family of apps and devices

New Image model announced too it looks like: https://ai.meta.com/research/publications/emu-enhancing-imag...

Man... building out a generic AI image generator, feels bad seeing pretty much all of big tech coming in.

mesmertechon Meta Quest 3

Standalone, but you can connect to PC if you want to play graphically intensive games. Quest 2 basically, you login with FB account so you need internet connection at first. Pretty "unhackable" afaik, you can install custom mods for games but thats pretty much it in terms of modifications I've seen.

mesmertechon Meta says its new speech-generating AI tool is too dangerous to release

Also there's this which offers a speedup to normal tortoise: https://github.com/152334H/tortoise-tts-fast

mesmertechon Meta says its new speech-generating AI tool is too dangerous to release

That sounds way too slow, just go with lowest or the second lowest quality. I think even the lowest quality works just as well.

Also you can rent a GPU on vastai or runpod, go with 3090 which should cost like 0.29 or even 0.15 if you go with spot pods.

mesmertechon Meta says its new speech-generating AI tool is too dangerous to release

https://github.com/neonbjb/tortoise-tts

Give this a try. You can run it locally if you have a good enough GPU. It is pretty slow at generation tho

mesmertechon Function calling and other API updates

I know, I meant it'd have been better if GPT4 got a slash but for my usecase 3.5 is still cost-efficient, basically would've been happy with either of them getting a price cut

mesmertechon Function calling and other API updates

Seems like I was right to wait for price drop on GPT3.5 a few months ago, was hoping for mostly a drastic drop for GPT4 but I guess 25%(for input tokens, effectively 12.5% on average) on 3.5 works as well

mesmertechon Ask HN: Who wants to be hired? (June 2023)

Location: Slovakia (Adaptable to different timezones)

Remote: Yes, Independent contractor

Willing to relocate: No, but open to travel

Technologies: 4 years in Python, Django, Flask, FastAPI, Celery, PostgreSQL, MongoDB, MySQL, AWS, Docker, Javascript, HTML, CSS

Experience with platforms like AWS, Azure, Digital Ocean, Hetzner, Stripe, Paypal

In addition: Worked with frontend technologies such as Nextjs, React, Vue, Zustand, Mantine, Vuex, Material UI, Ant-Design

Email: mesmerpercy@gmail.com

As an independent contractor, I'm seeking new challenges and opportunities. Open to starting with part-time engagements that may evolve into full-time roles. I have experience in managing small teams and have a strong understanding of UI and UX design.

We can hop on a quick video call to explore if we're a great fit for each other.

Resume/CV: On demand

Some extra info: I built and run a site that receives ~20 mill pageviews per month(private, can showcase on interview if needed). I'm currently interested in generative AI as side project, have experience creating serverless pipelines for GPU tasks to reduce costs instead of using something like Replicate, huggingface etc. Recent public examples of my work:

https://husks.org/ - Simple interface for ChatGPT api. Differentiating feature was document summarizing. Had other big plans for it but got too much competition

https://urpics.art/ - Built my own serverless api for the dreambooth training and inference to reduce costs. Currently working on a pipeline for spot pods to reduce costs even further(1/10th possible).

mesmertechon State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

Amazing model, close and probably better than Bard. Journey to getting the weights was a fun one : )

mesmertechon Block: Inflated user metrics and “frictionless” fraud facilitation

Why did paypal stock have to drop because of this...

mesmertechon Ask HN: Can any Hetzner user please explain your workflow on Hetzner?

- Deploy using docker swarm, CI ssh into machine, pull repo and run

- don't remember the last time I updated lol

- traefik + worker nodes on docker swarm

- again docker swarm

- I have a cronjob that makes backup using postgres, then uploads it to a digitalocean spaces, you can just use S3 as well

- I'm using cloudflare in front of server, but I also use inbuilt firewall as I host a postgres server with hetzner(only allow traffic from the web server worker nodes)

mesmertechon Google Bard Waitlist Parody

Not an accurate parody, it should say "Bard isn’t currently supported in your 'country'. Stay tuned!"

mesmertechon Bing Image Creator

So basically a wrapper for Dalle. More than the image thing the screenshots of "improved Bing" when you scroll to the bottom just look atrocious.

mesmertechon ChatGPT is down

Shamelessly plugging in my own "self-hosted ChatGPT UI" thing: https://husks.org/

mesmertechon Google Manipulated Digital Ad Prices and Hurt Publishers, per DOJ

A lot of concrete allegations compared to the search monopolization case I feel like. But with how slow these cases go, who knows how long it'll take to break up the ad server monopoly even if its successful

"With “Project Bell,” Google allegedly penalized publishers that did not give preferential access to AdX’s buyers, including Google Ads, by paying those publishers less revenues once an ad was placed. "

"if publishers tried to use a rival source of advertising demand for ‘first-look’ access to inventory, Google reduced bids by about 20 percent"

mesmertechon Microsoft 365 Down

Pretty much most of the azure services seem to be down, Azure Portal, Dev Azure, Azure functions, even the git repos aren't working

mesmertechon AI is in danger of being swallowed up by copyright law

It amazes me everytime, when it comes to AI images, ChatGPT there seems to be an aura of not giving a damn. But when it comes to code generation, the comments seem to turn into artists vs AI debate but for code.

mesmertechon Prisma laying off 28% staff

Been using Prisma quite a bit recently, wonder if this causes a slowdown in new features.

mesmertechon [dead]

I've seen several of these cost breakdowns for SaaS sites, but none for ad-supported websites.

Difference between the two is basically, ad supported sites need a lot of users that don't pay for anything but contribute by viewing ads where you don't need much marketing/ad spend. And SaaS most of your costs is going to be acquiring new users, as the service itself is very high profit margin.

Hope you find it helpful.