News for hackers. Stuff that matters.

ztratar

If you saw OpenAI launch Computer History you might have seen that data collection and aggregation are becoming more prevalent.

It turns out it's useful!

But having all my data sent to an AI lab in realtime feels wrong.

So I built an open-source version instead. Interestingly, local inference worked in my testing though it's certainly lower quality. Once we capture more data, we can explore LoRA adapters to improve local quality even more.

Open to feedback!

Re:Look, Another AI Browser on Wednesday October 22, 2025 @02:22PM (#45673138)
This is a funny post. Shows how deeply technical folks fail to understand business, risk, and open source.

Chromium is great. Why exactly should they innovate their first? A v1 should take whats available and not seek to reinvent the wheel.

Re:Show HN: LLM Based Spark Profiler on Wednesday April 09, 2025 @08:03PM (#43639296)
Also curious how the agent works?
The worst ever YC demo and company I've ever seen.

Technology is supposed to make the world better. This is vile, and the script itself shows the founders do not care or understand what they are doing.

Love the idea.
I guess, just my opinion (probably worthless), if you can get it to speed up even ~20%, I think that would be enough to make the average person agree.
How much faster will the robotic arms be able to go? Currently this looks far too slow, though as a v1 it's great.
Re:Sora is here on Monday December 09, 2024 @02:22PM (#42369503)
Those look... far worse? What am I missing.
Re:Building Databases over a Weekend on Wednesday November 20, 2024 @11:55PM (#42201254)
I almost thought the opposite, but im no db guy.
Re:Simone giertz talks about invention on Tuesday October 22, 2024 @08:06PM (#41920072)
But annoyingly sometimes this is great advice. It's not that dissimilar to YC's make something people want, but it relies on you having taste/needs/wants that are an addressable market. Which is sometimes true.
Will be excited to see the typescript bindings once out. We may be able to use this to handle some of our workloads at Embra.

Will reach out! Congrats on the ship.

Re:Starlink Mini on Friday July 12, 2024 @02:08PM (#40947960)
The difference in the size of the TAM of consumer vs military application must be wild.

Seems radically useful for military apps. Somewhat useful for RV/camper vans. Cool either way.

Sounds to me like he made a couple unfortunate moves:

- Propose a 5 year plan (lol) - Don't lead the incident with leadership, but with blame - Speaking about problems more than solving problems (hard to do) - Lack of relationship building

On one hand, I empathize with Chris. On the other hand, this sounds to me like he just didn't know how to perform in this environment. And that's totally fine! Not everything should learn how to perform in the bureaucratic knots -- startups are simpler in this way. And there's a reason the big guys lose their edge over time, and then some exec in a board room is faced with -10% YoY loss without any truthful VPs around the table.

Re:NotebookLM: An AI Notebook on Wednesday July 12, 2023 @03:03PM (#36700037)
For such a good idea, this design is comically bad.
Re:Native JSON Output from GPT-4 on Wednesday June 14, 2023 @05:35PM (#36333084)
Nah, this was solved by most teams a while ago.
Re:Microsoft is integrating Copilot into Windows 11 on Wednesday May 24, 2023 @12:15PM (#36060233)
If you're interested, we're building Co-pilot for Mac. https://embra.app

:)

Show HN: Embra Plugins

https://embra.app/plugins

Hey HN! We just announced Embra Plugins, a developer integration path similar to ChatGPT plugins that works with our desktop client.Embra works with multiple models, and can eventually integrate local models as well. This allows our plugin ecosystem to eventually run entirely on your machine – privacy preserving!Given OpenAI allowed plugins to follow the OpenAPI specification, this means anyone...

( Read More... | no comments | on HN )
Given the model performance is thus affected by a k-nearest neighbor, but those algorithms are proving not great for baseline vector search, how well will this actually work?

It seems mostly like a vertically integrated vector DB + existing LLM call, but correct me if I'm wrong. There are of course some performance gains with that, but the holy grail of "understanding" at unlimited length still seems unsolved.

Re:Stripe Faces $3.5B Tax Bill as Employees' Shares Expire on Wednesday March 08, 2023 @11:56PM (#35078439)
Stripe moved off of First Data and become its own processor way back in 2019.
Re:Stripe Faces $3.5B Tax Bill as Employees' Shares Expire on Wednesday March 08, 2023 @11:55PM (#35078436)
And those are measured in the same way Stripe measures the $14B number.

Granted, PayPal has a higher take rate than both Adyen and Stripe. But also much lower growth.

Re:Stripe Faces $3.5B Tax Bill as Employees' Shares Expire on Wednesday March 08, 2023 @08:26PM (#35077224)
Tax liabilities on (optional) employee secondary stock sales is most certainly not a normal business operation.
Re:LangChain TypeScript Support on Friday February 17, 2023 @04:21PM (#34840733)
Oof. Nevermind. i don't seem streaming in the new docs.

I need 100% of my output streamed. I rarely ever have a use case that doesn't include streaming.

Re:LangChain TypeScript Support on Friday February 17, 2023 @04:15PM (#34840675)
Literally struggling through python right now, and this is amazing.

Helllllllll yeah.

Given these are fines from the card networks, the business, if it's in the right, can go to court quite readily.

Illegal processing is no joke. This isn't as much "Stripe ruining a business" as it is "a business appears to have been processing illegal activity, which puts it into liability".

To be clear, in some jurisdictions and depending on the activity, illegal processing also carries a jail sentence! The most extreme of which is, of course, terrorist financing.

Thanks! Your document tool looks great as well.

Might build more extensive file support, if users ask for it.

Show HN: Embra – A fast, Chat-GPT like assistant for your Mac. App integrated

Hey HN!We're building Embra, an always-one-second away GPT assistant that integrates into your OS, apps, and workflows.Everyone is talking about the "context problem" as it relates to GPT, in that the answers you get out are only as good as the information you put in. And, well, a lot of users don't want to type in a bunch of text.With Embra, you can pick and choose which context to feed the AI...

( Read More... | 5 comments | on HN )
Re:Chronophoto on Saturday January 28, 2023 @03:11PM (#34561274)
This was a lot of fun. Very unique game. Congratulations on making it! :)
Re:Tesla AI Day 2022 on Saturday October 01, 2022 @06:45PM (#33050877)
If you're in tech and worried about job losses, you may want to look at how history has literally done this 100 times in the past to numerous industries.

Bringing race into this is nonsense. If the robots are taking jobs, do you think they care what color someone is? They don't. And it's literally going to come for everyone's current job. Which is a good thing, so long as we have an economic system that evolves alongside it. This is quite literally the only way to eliminate poverty in totality.

Older »