HN user
Well, it’s up to the user or post-trainer of the LLM what they believe to be above average. Then they can design around that.
In the case of real world LLMs and post-training, what is above average is defined roughly as: labeled good by expert humans, and scoring high on RL environments related to coding like debugging, passing tests, or running efficiently and verifiably correctly.
LLMs learn a distribution during pre-training, not only an average.
Then, by giving them context or by post-training, you can make them sample non-average parts of the distribution they learned.
I checked out your agent and it looks pretty well designed. Congrats on starting to share it with others!
One thing I noticed: "Your Tools: Aether agents get tools exclusively via MCP servers." "...Aether ships with 1st-party MCPs for file system operations..."
Can you share your thoughts on why you decided to use MCP as the core tool abstraction? I have heard many decry MCP as being context-wasteful. Is this not the case with your agent?
Why does its architecture or you knowing how AI is architected cause thoughts of it being conscious to go out the window?
It seems like the biggest factor has nothing to do with AI, but instead that you went from being someone who admits they don’t know how consciousness works to being someone who thinks they know how consciousness works now and can make confident assertions about it.
That’s the most common time control on chess.com according to this article: https://www.chess.com/article/view/time-controls-and-game-re...
It sounds like a good idea to me!
Hey! I played against a bot and it was pretty fun.
Small suggestion: too many queues can make it very difficult to build up a network of players at first. I'd suggest, for now, lowering the amount of available time control queues so that two players who happen to be on at the same time are more likely to actually find a game.
AI comments are against Hacker News rules.
The sheer em dash density of this post really struck me, so I asked Claude to write a script which ranked text post Show HNs over the last week in order of em dash density. Script here: https://github.com/mturnshek/hn-em-dash-density/tree/master
This post comes in 12th place out of 668 with 0.6232% em dash density. I was also surprised by the large number of ShowHNs in the last week.
Here are the highest density 4 from the last week:
1. 1.2195% — Show HN: AI-native memory, recall and reminder on CLI – 100% local with Ollama (https://news.ycombinator.com/item?id=47417498)
2. 1.0417% — Show HN: CEL v0.2 Pro – cryptographic black box recorder for AI systems (Python) (https://news.ycombinator.com/item?id=47406917)
3. 1.0174% — Show HN: FalconAI – AI-Powered Smart TV Streaming, Search and Voice Control (https://news.ycombinator.com/item?id=47377118)
4. 0.9259% — Show HN: Ostov.js – Backbone.js Fork Without jQuery/Underscore, Classes, TS, ES (https://news.ycombinator.com/item?id=47401800)
Posts with at least one em or en dash: 457/668 (68.4%)
I also learned about the existence of the en dash. I only knew about the em dash and the hyphen, but there's:
— em dash (U+2014)
– en dash (U+2013)
- hyphen-minus (U+002D)
Anyway.
Are you trying to imply that humans don’t need generalized knowledge, or that we’re not “rewarded” for having highly generalized knowledge?
If so, good luck walking to your kitchen this morning, knowing how to breathe, etc.
that corporate profits would rise while consumer spend dropped are literally incompatible realities
These are not incompatible realities.
I would be willing to accept the statement that corporate revenues increasing and consumer spending decreasing are incompatible realities.
But it’s feasible to think the following occurs:
- labor income falls
- consumer spending drops
- corporate revenues drop
- corporate profits moderately increase because profit margins get much higher
- government deficit continues (which, from an accounting perspective, means other accounts are in surplus, potentially US corporations)
I’m not saying I strongly predict the above, necessarily! I just don’t think it’s correct to say it’s not a conceivable reality.
When I hear "coding agent", I think of both the harness and the LLM as a pair. Like, Claude Opus 4.6 and Claude Code is a coding agent, or Gemini 3 Pro and Pi is a coding agent.
"Harness" is a way to reference the coding agent minus the "LLM" part.
If an agent is an LLM in a loop with tool calls, there are two components: 1) the LLM. 2) The loop with tool calls. That second part could be called the harness.
LLMs are not "average text generation machines" once they have context. LLMs learn a distribution.
The moment you start the prompt with "You are an interactive CLI tool that helps users with software engineering at the level of a veteran expert" you have biased the LLM such that the tokens it produces are from a very non-average part of the distribution it's modeling.
It is not a "narrative", "philosophical paradigm", or him "getting high on his own supply". It is simply him sharing his thoughts about something.
It really doesn’t, at all. Every sentence has a clear, non-equivocative meaning and it doesn’t use any LLM tropes. Your LLM sensor is seriously faulty.
What is the goal of doing that vs using L2 loss?
So, I have heard a number of people say this, and I feel like I'm the person in your conversations saying it's a coarse description and downplays the details. What I don't understand is, what specifically do we gain from thinking of it as a Markov chain.
Like, what is one insight beyond that LLMs are Markov chains that you've derived from thinking of LLMs as Markov chains? I'm genuinely very curious.
Boglehead
140% gain on your holdings this year
Choose one.
Why would one have motivation to not use activation functions?
To my knowledge they’re a negligible portion of the total compute during training or inference and work well to provide non-linearity.
Very open to learning more.
Okay, that makes perfect sense. Thanks for the explanation!
Can you please explain the insight about reduced workweeks you are deriving from what you've linked? It is not obvious to me.
we’d just have to do it
Highly economically disincentivized collective actions like “pulling the plug on AI” are among the most non-trivial of problems.
Using the word “just” here hand waves the crux.
Can you list some long textbooks on a single subject that are amazing?
Don’t forget the training data!
The phrasing “<statement> — no X, Y, Z, just <final simplified claim>” is cropping up a lot lately.
4o also ends many of its messages that way. It has to be related.
Dozens? Don't you mean, probably hundreds of trillions in the observable universe? Not that the number of planets really implies anything when we don't know the probability of life arising on one of them.
There are many effective ways to argue against what you're saying, but I'll choose the easiest.
You said "perhaps" it doesn't matter. Well, perhaps it does matter. If there is even a small chance it does matter, we might as well act like it matters.
So, your argument (ignoring the appeal to authority) is:
1.Social rights can be reduced entirely to obligations on others.
Dubious. Ignoring that maybe the value to the person granted the social right could outweigh the obligation to others, or that the holistic effect on the system might be greater than its reduced parts.
2. Social rights need to be enforced by those with power.
3. Supporting and legitimizing social rights means supporting and legitimizing a specific source of power that will enforce them.
The jump from 2 to 3 is not entirely clear to me. I think it's a valid stance to be critical, and even unsupportive of any given source of power enforcing social rights, while being fully supportive of the right itself.
4. The ones who end up in positions of power are the most ruthless and power-hungry.
5. Those people being in positions of power is so bad that it outweighs any benefit from the original cause, 1 + 2.
Overall, I'd say the critical flaw of your argument is that it's too reductive and assumes that everything very neatly follows linear, simple paths of power.
In addition, I think you can replace the "social rights" in point 2 with any law, and you will have the same points 3, 4, and 5.
How is “losing weight is simple: eat fewer calories than you burn” unhelpful? I’m genuinely asking, and want to understand how to empathize better.
It seems like a lot of people don’t make weight loss progress because they never internalize that fact and seem to “fight it” psychologically.
When and how would one use binary vectors for encoding in ML? Do you have to make your model work natively with binary vectors or is there a translation step between float and binary vectors to make it compatible?
What went into creating the color scheme?