HNHacker News
TopNewBestAskShowJobs

seunosewa

1,883 karma · joined April 10, 2007

submissionscomments
seunosewa··on Some conservationists are helping to restore Africa’s wild dog populations
They are cute but nasty. They eat their prey alive in the most painful way.
seunosewa··on Xiaomi: New CPU matches Apple cores single threaded, much faster multithreaded
From the benchmarks, the single core performance is close but lower.
seunosewa··on Nvidia Nemotron 3.5 Lightning and NeMo Switchyard
Unfortunately?
seunosewa··on Claude Code May–August 2026 weekly limits promotion
$100 Claude and $100 ChatGPT Pro is the best value for $200/month. They can see each others' mistakes.
seunosewa··on MathCode, Mathematical Coding Agent
Could you provide a practical example?
seunosewa··on Gemini 3.7 Flash
Last time I tried it, Flash introduced too many errors due to sloppiness. Is it more reliable at following instructions now?
seunosewa··on Gemini 3.7 Flash
They want to maintain the perception that Flash is worth $7.5/mot, so they can charge more for the next one.
seunosewa··on Gemini 3.7 Flash
Two thoughts:

1) It makes sense to try 3.7 flash before cancelling.

2) Prompting models to be honest is surprisingly effective in my recent experience. But only if they listen to instructions.

seunosewa··on DeepSeek V4 Pro 0813
Could you try setting the temperature very low e.g. 0.0?
seunosewa··on DeepSeek V4 Pro 0813
You can set it up to always use the official provider.
seunosewa··on DeepSeek V4 Pro 0813
Do it a second time at least.
seunosewa··on Antigravity 2.0
The SSH into machine market is huge. Cloud work, Dedicated servers, VPS, Colab.
seunosewa··on Elevated errors on Claude Opus 5
GPT 5.6 Sol is so free of drama for me for brainstorming, writing prompts, etc. Tasks I always did with Opus until 4.7.
seunosewa··on Claude Opus 5
Provide an option. An autonomous mode or an interactive mode.
seunosewa··on Claude Opus 5
The problem is it should ask you before doing these things. This is a cherry picked example. There are times when they go on misadventures.
seunosewa··on Qwen 3.8
They are trying to make money. That's what firms in any capitalistic economy care about the most. Regardless of the government's presumed interference, the companies themselves are all trying to make money. All competing for subscriptions and API payments.

One aspect of this is making a name for yourself i.e. PR. Making a capable model open source helps a lot with that.

seunosewa··on Grok 4.5
How does Fable compare?
seunosewa··on Grok 4.5
The input price is quite high. That's what gets you.
seunosewa··on Grok 4.5
Not true. The only issue is cost of frontier models.
seunosewa··on The short leash AI coding method for beating Fable
One of my favourite things to do is to ask the models, "what does this code do and why?" They are usually not far from the truth. From the perspective of a LLM that has been trained on all of the public code on GitHub, issues, and pull requests, your code and its intents are not that unique. They can see patterns in the code that point to the purpose. The constraints that shaped your code are the same constraints everyone else had.

Developer intents and aspirations aren't unique either. And in an LLM-written codebase, the chat history and prompts are available. The constraints, the rejected approaches, the goals. Users ask the models to write MD files that directly state their intents and aspirations for the ode.

seunosewa··on The short leash AI coding method for beating Fable
We often don't understand the code we wrote 6 months ago.
seunosewa··on The short leash AI coding method for beating Fable
You can just ask the model to explain the code to you.
seunosewa··on Qwen 3.6 27B is the sweet spot for local development
You can get some work done by using low power mode even when plugged in, and making your fan start running when the temps just start to rise (maybe 40 degrees. Use a third party fan app to set it up
seunosewa··on VibeThinker: 3B param model that beats Opus 4.5 on reasoning with novel SFT+GRPO
Then smaller the models are, the longer they have to reason when dealing with complex problems. The trade-off is real.
seunosewa··on We should be more tired than the model
Unless you log its mistakes and how they were solved in decisions.log
seunosewa··on YouTube to automatically label AI-generated videos
AI has completely ruined animal short videos on Youtube. Videos of pets behaving like humans are everywhere. At first they warm your heart, then you realize that you've been tricked.
seunosewa··on If AI writes your code, why use Python?
What do you mean? Those are valid identifiers but programmers aren't required to use them.
seunosewa··on If AI writes your code, why use Python?
Less verbose languages also use fewer tokens, saving precious context.
seunosewa··on Grok 4.3
They dropped the output cost, butthe input cost is relatively high. This is a recent trend. Seen with DeepSeek 4 Pro as well.
seunosewa··on Qwen3.6-Max-Preview: Smarter, Sharper, Still Evolving
API costs are really high compared to subs.
← PreviousPage 2 of 33Next →