augusto-moura
Joined 9 years ago
Karma
1,284
Submissions
456
Member since
Apr 22
Brazilian developer, interested in hacking anything, from container orchestration to breadboards. I also know a lot of JavaScript
View on Hacker NewsRecent submissions
Comment · 6 days ago
They were never useful for the results, just for market expectation, which is never the same as real chances of something happening
Comment · 6 days ago
Why was this flagged? It seems relevant
Comment · 23 days ago
Similar in Brazil, orkut was the social network here
Comment · 25 days ago
It doesn't, it is a runtime for running Portable Executables[1] on the fly while also mocking windows API libraries to Linux implementations
Comment · 25 days ago
Nice moment to hook the git-maintenance[1] command, which setups automatic pre-fetching of a repository in a scheduled manner. That way you don't need to remember to manually update your mirrors. Since the repo is public, cloning with https is a simpler setup for prefetching, but ssh is also supported with some configuring.
Edit: wrong link
Comment · 1 month ago
Not only training, you can literally clone the source code locally and ask the agent to debug your problem by grepping the kernel itself. Even if doesn't "know" the answer it can investigate it on the fly, that's super powerful
Comment · 1 month ago
No auto mode is the thing that bothers me, I don't trust the cli blindly nor do I trust myself to read every python script it throws at me. Auto mode is an acceptable middle ground in my experience
Comment · 1 month ago
Most people say full agentic coding only got good enough for production at end of last year.
This year was the turning point for most companies. A lot of experiments going now right now.
Of course things can change drastically in the next years, as they already have in the last few.
Only thing I know is that it is really stressful working on this rat race
Comment · 2 months ago
Have you tried setting up systemd-oomd [1]? It works by killing the offending process (hopefully) before a true OOM.
When I run in a frozen system because of RAM, I'm a happy user of the Magic SysRq 'f' [2], it bypasses most kernel scheduling so it works even on frozen systems
Comment · 2 months ago
For OOM, there's earlyoom [1] and systemd-oomd [2]. They usually work out of the box
Comment · 2 months ago
Like xkill? [1]
Just be careful where you click
Comment · 2 months ago
IMO AI companies are selling the infrastructure and easier access to the models, beyond competing to get the smarter model. We can run Qwen locally, but at what cost? A good enough GPU for running 27B is more than 4,000 USD where I live (Brazil). Even if I were to spend 100s of hours in credits I would never reach a point where the 4000 USD investment pays off. Unless I'm doing it for the sport, building my own LLM rig is not financially viable.
This is without saying that company clients of Anthropic or OpenAI will not have the upfront money to get a RTX 3090 for each and every employee. Much better to sign a deal with the LLM providers and get cheaper token price with a sales person.
Comment · 2 months ago
If the public interest is divided, what should be ICE's interest? This applies to other enforcing orgs too. Should it take the majority even if the debate is close to a tie?
In most cases, what wins is the interests of people in power, for all domestic policing
Comment · 2 months ago
Would they not allow it? Rockstar is the owner of the game, they can update any removal of content if they want to. Modern games frequently change gameplay, story, and soundtrack between versions
Comment · 2 months ago
You can dowgrade a version of a game, if you want to. Is not as easy as clicking a button, but it is possible and not against ToS