seunosewa

8 ms·

91.
by seunosewa 1y ago
Frontier models aren't marginally more costly. They are 6 to 10 times as expensive based on their API pricing.
92.
by seunosewa 1y ago
I wonder how it compares the GPT 5.1 Mini and Gemini 2.5 Flash.
93.
by seunosewa 1y ago
The LLM objective is whatever they are trained to do, whether it's completing text, obeying instructions, coding, etc. In pre-training, we drop a lot of human-written text in them. This allows them to learn the rules of language and gr
94.
by seunosewa 1y ago
And this model is arguably their least impressive model.
95.
by seunosewa 1y ago
We do the same thing. We pick words that are statistically likely to get us what we want. And much of it is unconscious. You don't formally reason about every word you speak. You are focused on your objective, and your brain fills in t
96.
by seunosewa 1y ago
Claude Code is free if you have a subscription to Claude. It's much more affordable than using the API key directly.
97.
by seunosewa 1y ago
If you need precision, that's the way to go, and it's usually cheaper and faster too.
98.
by seunosewa 1y ago
Sometimes it will randomly generate something like this in the body of the text: ``` <tool_call>executeshell <arg_key>command</arg_key> <arg_value>echo "" >> novels/AI_Voodoo_Romance/ch
99.
by seunosewa 1y ago
The DeepSeek R1 in that list is the old model that's been replaced. Update: Understood.
100.
by seunosewa 1y ago
It's a hybrid reasoning model. It's good with tool calls and doesn't think too much about everything, but it regularly uses outdated tool formats randomly instead of the standard JSON format. I guess the V3 training set has a
101.
by seunosewa 1y ago
The first self-announced vibe coder was a founder of OpenAI. So he had the knowledge. He started to trust the code written by the models to a point where he didn't always read the code. This happens when you you have learned the exact
102.
by seunosewa 1y ago
When prompt engineering isn't giving you reliable results.
103.
by seunosewa 1y ago
Claude is expensive. Falling back to a more expensive model seems counterproductive.
104.
by seunosewa 1y ago
You can choose to write the juiciest parts yourself and let the AI deal with grunt work.
105.
by seunosewa 1y ago
You should try GLM 4.5; it's better in practice than Kimi K2 and Qwen3 Coder, but it's not getting much hype.
106.
by seunosewa 1y ago
Also great.
107.
by seunosewa 1y ago
They will train themselves by doing open source projects with AI.
108.
by seunosewa 1y ago
Providing agents with relevant documentation including code samples and API references should help a lot in your scenario. Relevant documentation helps a lot when you're working with obscure languages or libraries.
109.
by seunosewa 1y ago
It's ridiculously overpriced in the API. Just like o3 used to be.
110.
by seunosewa 1y ago
GitHub made git popular, and git made GitHub popular. So the major mercurial hosts (especially bucket) decided that it wasn't worth supporting even for their existing users. That was the mail in Mercurial's coffin. Google also dec
111.
by seunosewa 1y ago
What are you using to fire up requests on your phone?
112.
by seunosewa 1y ago
What did the LLMs get wrong most frequently when working with C#?
113.
by seunosewa 1y ago
Printing stack traces generates a lot of useful context but it's not done enough.
114.
by seunosewa 1y ago
The Cerebras.ai plan offers a flat fee of $50 or $200. The API price is not a reason to reject the subscription price.
115.
by seunosewa 1y ago
They search for the token of interest, e.g. grep -n then they read that line and the next 50 lines or so. They continue until they get to the end.
116.
by seunosewa 1y ago
Yes, of course. The pretence of ignorance in the article is hilarious.
117.
by seunosewa 1y ago
Nope.
118.
by seunosewa 1y ago
"Exception-Safe¶ There are no runtime exceptions. All exceptions have defined semantics, and they are always caught." So checked Exceptions like Java?
119.
by seunosewa 1y ago
I loved his wacky songs. He made poisoning pigeons in the park and nuclear proliferation funny.
120.
by seunosewa 1y ago
The LLMs can easily be trained to teach people how to use new programming languages.

More ›