something I've noticed is that on those days when I spend the majority of my time vibing with projects that don't matter and will only be used one time, such as this integration I've built now, I am exhausted.
my brain is melting after a day of this LLM usage
Good morning you all
Started off by vibe-coding a change to an internal tool. Now time for some rust
Anyone working on something fun?
So, @bagder.mastodon.social.ap.brid.gy I guess you might be a bit skeptical of this 😅
I get what Garry mean when he produce this types of code :D
Garry Tan, the CEO of Y Combinator said they write 10k lines of code.
I asked Opus 4.5 to write a solution to encode only the filename in a url. The first image is AI and the second is my solution.
The AI code got bugs, for instance if we encode a char outside of S3_ENCODE_STR.
We bought computers for the company and I want the mac hardware so I made a switched :D But my personal computer is still arch
How company c-suites look at AI when their layoffs is now AI innovation
Good morning you all. Today we do working class coffee and debugging a error
Today will be a day of fixing bugs
Product managers when devs say "it's not technically possible" for the 3 time this week
Yeah, it's mostly about the amount of packages that exist. It's easier to get something up and running in GO because there already exists tons of libraries.
I also dislike opam and how opam works but dune fixes all the problems I have with opam.
I love ocaml but the problem with ocaml is the ecosystem. It has improved drastically but it's still far behind because if we want it to improve more do we need more developers and FP programming languages in general have quite few devs outside research
Refactoring...... Refactoring...... Refactoring......
Vibe coders will look at this solution and think, well, this look good to me
Tested if opus could figure out a deadlock issue :)))))
but coffee in the morning is a no brainer
evolving
Curved ultrawide monitor and laptop showing code and apps, mechanical keyboard and gaming mouse on a large desk mat, headphones, soda can, GitLab notebook.
At the moment, I'm developing a new feature for a side project focused on data analysis and automation to uncover synonyms. The goal is to leverage real user search data to determine if suggesting synonyms could enhance search results.
Stack is #Rust, Meilisearch & Postgres.
This is the reason why RAM cost 900$ btw
The hardest part about this type of service is not building the app that talks to other systems. Is all the security around it such as firewalls or analysis of traffic that cost much money and is really hard to get secure. If this was easy would Anthropic or Openai 100% already built a version
But it's also horrible and dangerous solution because we now have an LLM with capacity of opening and reading apps, browsing the web and even executing CLI commands, and we run it in an unsupervised environment with no human in the loop where anyone can prompt it via WhatsApp, Discord and emails
I get Openclawd is popular because is the closes solution to the demand of having your own personal assistant
Increasing the salary
Now I am double checking the payload we send so that we send it the way we want
I also need to take some time and improve my work desk at the office, It's about time ☺️
Working on a integration between us another service, I build the foundation for the integration such as the http client, how to fetch data and so on.
Then I let @ampcode.bsky.social build the 4 other requests we want to make against a swagger I got. Worked like a charm
Peak
It doesn’t necessarily mean curl will be less safe overnight, but it definitely means that talented people won’t be looking for exploits in it as much.
With the recent changes, curl will eventually miss important exploits because:
1. AI "slop" reports are taking up most of the maintainers' time.
2. The removal of monetary rewards takes away one of the main reasons talented exploit researchers look for vulnerabilities in curl in the first place.
