Hacker News: carlosbaraza

New comment by carlosbaraza in "The United States and Israel have launched a major attack on Iran"

carlosbaraza — Sat, 28 Feb 2026 08:22:35 +0000

What are that pizza place google statistics?

New comment by carlosbaraza in "Spain’s LaLiga has blocked access to freedom.gov"

carlosbaraza — Mon, 23 Feb 2026 23:34:13 +0000

Cloudflare offers some genuinely valuable services that protect you from exposing your infrastructure to the world wild web. And regardless why does a private institution like LaLiga have the power to censor anything they want for their own benefit?

New comment by carlosbaraza in "Spain’s LaLiga has blocked access to freedom.gov"

carlosbaraza — Mon, 23 Feb 2026 23:30:57 +0000

My point is that people don't understand nor know that this is happening at all. Even when I get customers complaining that the service isn't available, they don't believe that their ISP is blocking them because of football. It's almost unbelievable for how absurd it sounds that people don't even think that could be the reason.

New comment by carlosbaraza in "Spain’s LaLiga has blocked access to freedom.gov"

carlosbaraza — Mon, 23 Feb 2026 07:12:17 +0000

I have commented this in multiple occasions. What is happening here in Spain with LaLiga is just absurd. My company's domain gets blocked often because we use CloudFlare. In essence, any service using CloudFlare gets blocked often. The main problem is that the common Joe tries to navigate and finds that it doesn't work, and they blame their network, and when they come back two hours later after the game finished, the website works, so they move on. The only way for this to get resolved is if they blocked something critical and an accident happened because of that (e.g. hospital services, traffic control, or something like that). Eventually this will escalate to national courts (currently this was dictated by a regional court in Barcelona). But again, legal action is extremely slow. VPNs are becoming a must everywhere, because the Internet is becoming wild from all directions.

New comment by carlosbaraza in "Cowork: Claude Code for the rest of your work"

carlosbaraza — Thu, 15 Jan 2026 08:41:36 +0000

some things will never change :)

New comment by carlosbaraza in "Cowork: Claude Code for the rest of your work"

carlosbaraza — Tue, 13 Jan 2026 13:28:42 +0000

Same thing for me. It crashes. Submitted a report with the "Send to Apple" report, not sure if there is any way the team can retrieve these reports.

New comment by carlosbaraza in "Composer: Building a fast frontier model with RL"

carlosbaraza — Wed, 29 Oct 2025 18:07:49 +0000

I have never leveraged git worktrees... That is such a crazy useful tool that I am almost ashamed of not having researched it before. Git is such a beautiful piece of software.

New comment by carlosbaraza in "Composer: Building a fast frontier model with RL"

carlosbaraza — Wed, 29 Oct 2025 18:02:10 +0000

How do you work with multiple agents?

New comment by carlosbaraza in "Composer: Building a fast frontier model with RL"

carlosbaraza — Wed, 29 Oct 2025 18:00:38 +0000

This is nice. I liked Cheetah for grunt work that I want to get out quickly and is not too hard. The speed is really awesome. A model that would run at even higher speeds like the OSS models at groq/cerebras would really be workflow changing, because the slowness of SOTA models really breaks the flow. I find myself taking a ton of breaks and getting distracted while I wait for a model to complete a task (e.g. just now).

New comment by carlosbaraza in "Composer: Building a fast frontier model with RL"

carlosbaraza — Wed, 29 Oct 2025 17:50:34 +0000

Cursor 2.0 keeps crashing on me while having an agent running and opening the IDE part of the application. I might have to rollback.

New comment by carlosbaraza in "Composer: Building a fast frontier model with RL"

carlosbaraza — Wed, 29 Oct 2025 17:47:05 +0000

Could anyone explain how to use multiple agents and subagents in Cursor, Claude Code, or others? It is already challenging to me taming one model doing work, let alone synchronizing multiple parallel workers.

Do you have to split the plan in parallelizable tasks that could be worked in parallel in one codebase without breaking and confusing the other agents?

New comment by carlosbaraza in "LaLiga's Anti-Piracy Crackdown Triggers Widespread Internet Disruptions in Spain"

carlosbaraza — Mon, 22 Sep 2025 05:55:35 +0000

I live in Spain. The main problem is that your internet becomes patchy and things stop working for a few hours during the match. People notice the incident, but they don't understand what is happening, they just move on and come back few hours later for it to be resolved. Only some people understand the depth–usually tech people like us that get paged–and complain, and then the president from LaLiga call us a "bunch of freaks".

I have personally sent letters to everyone, including the court that ruled and allowed this whole mess. The judge in question must be either extremely incompetent or corrupted. The court response to this problem wasn't acknowledging it, but to double down, because it seems like some egos were hurt.

The only way I can imagine this situation stopping is by someone dying from a core healthcare system malfunctioning and a court case becoming viral. In the meantime, a single private institution can destroy the internet whenever they want, with full legal backup. Just insane.

New comment by carlosbaraza in "Gemini 2.5 Flash Image"

carlosbaraza — Wed, 27 Aug 2025 07:21:08 +0000

Basic things like: "{uploaded image of a man} can you remove the glasses?" or "make everyone in the picture smile" or "open the eyes of everyone in the photo". Nothing that a human would consider "unsafe". I am based in EU and using Google AI Studio with all safety toggles set to "Off".

New comment by carlosbaraza in "Gemini 2.5 Flash Image"

carlosbaraza — Tue, 26 Aug 2025 21:32:00 +0000

Unfortunately, it suffers from the same safetyism than other many releases. Half of the prompts get rejected. How can you have character consistency if the model is forbidden from editing any human. And most of my photo editing involves humans, so basically this is just a useless product. I get that Google doesn't want to be responsible for deep fake advances, but that seems inevitable, so this is just slightly delaying progress. Eventually we will have to face it and allow for society to adapt.

This trend of tools that point a finger at you and set guardrails is quite frustrating. We might need a new OSS movement to regain our freedom.

New comment by carlosbaraza in "Cloudflare CEO: Football piracy blocks will claim lives"

carlosbaraza — Mon, 26 May 2025 20:37:51 +0000

I live in Spain and my ISP is Digi, which uses the network from Telefonica. These blocks are incredibly frustrating, and a ton of people have noticed websites and services not working. However, because the block lasts some hours, people don't know what is happening: "is my mobile network bad?", "Is the website down?". They try a few hours later and it's back up, so they move on.

My company's website is behind Cloudflare and I discovered this whole situation because someone couldn't access it. Also my home assistant is not accessible from the internet the days with a match. And we use it to open the garage and the house. We learned the lesson the hard way being locked outside until I managed to connect with a VPN. This is just nuts and incredibly frustrating. And for La Liga we are just a bunch of "frikis" (nerds) complaining about it... because we are the only ones that understand what the problem is.

Unfortunately, someone would have to die and a lawsuit to follow, and maybe that could stop this crazy nonsense. E.g. A few days ago I read about someone with diabetes whose device was malfunctioning because of these blocks.

New comment by carlosbaraza in "Improvements to data analysis in ChatGPT"

carlosbaraza — Fri, 17 May 2024 10:46:03 +0000

One interesting point of this feature is that it acknowledges that natural language is not a sufficiently good interface, we still need custom UI for particular use cases. As long as humans are in the driver seat, multiple UIs for the same problem will be created and used by many, enabling the SaaS ecosystem to thrive with this new LLM tooling backend.

New comment by carlosbaraza in "GPT-4o's Memory Breakthrough – Needle in a Needlestack"

carlosbaraza — Wed, 15 May 2024 09:46:38 +0000

I have the same feeling. I asked to find duplicates in a list of 6k items and it basically hallucinated the entire answer multiple times. Some times it finds some, but it interlaces the duplicates with other hallucinated items. I wasn't expecting it to get it right, cause I think this task is challenging with a fixed amount of attention heads. However, the answer seems much worse than Claude Opus or GPT-4.

New comment by carlosbaraza in "Show HN: I built a non-linear UI for ChatGPT"

carlosbaraza — Thu, 09 May 2024 08:41:43 +0000

Sometime ago I had an idea for a similar interface without the dragging feature. Basically, just a tree visualisation. I usually discuss a tangent topic in the same conversation, but I don't want to confuse the AI afterwards, so I edit a previous message when the tangent started. However, OpenAI would discard that tangent tree, instead it would be nice to have a tree of the tangent topics explored, without necessarily having to sort them manually, just visualising the tree.

New comment by carlosbaraza in "Unexpected responses from ChatGPT: Incident Report"

carlosbaraza — Thu, 22 Feb 2024 08:02:01 +0000

If a human had this failure, it would probably be something like a psychotic episode. If a super intelligence had a psychotic episode because of a bug, it could be pretty destructive.

New comment by carlosbaraza in "Ask HN: Who else is working on nothing?"

carlosbaraza — Sun, 14 Jan 2024 10:05:05 +0000

I feel the same way. I follow the LLM/AI advancements very closely, and I find that motivating. However, the whole discussion seems to turn around long term replacement of all labor and absorption of business value into some big tech providing the AI infrastructure.

I usually had two types of motivation, technical motivation to learn skills that would eventually be useful in my career and hacker motivation to build something with the implicit expectation of eventually making it a sustainable business. But now I am not certain there is a point in learning a new skill or building a side hustle if eventually an LLM will be better than me or my business for pennies.

Maybe this whole feeling is part of the AI replacement hype, and totally false. And maybe it is just common to feel this way when technological revolutions happen, and it's just temporary and we should push through. Regardless, we probably should keep monitoring the space.

Maybe it's time to genuinely focus on doing anything that intrinsically makes you feel good without any external expectations. An LLM might be better than you at everything, but you would still enjoy your mediocre implementation, just because you did it. Maybe this is what it really means to be an artisan.