HN user

b1n

376 karma
Posts12
Comments77
View on HN
Qwen3-Max-Thinking 6 months ago

This answer is the most important one.

The future of state LLMs is not censoring subjects - it's slowly but surely persuading people using your LLM that your version of events - or your spin on that event - is the truth.

I've invented a replacement for the PDF format.

It's called Locked Markdown. `*.lmd`.

It provides 98% of the functionality most people use PDFs for, without all the extra bullshit.

The Spec:

- The most popular Markdown Spec (git flavour?)

- At the top of the page place a hash sum of the markdown below a marker on the page.

LMD Reader:

- Prevents editing the markdown.

- Warns user if the hash sum is not valid.

Example:

```lmd

sha256 77ec0f678315f8a207c3501137e1dfc9642b79a9c93e21807df7b5242846c05c

-------------------------------

# Header 1

Paragraph of text about how bloated the PDF format is.

```

Zip .lmd together with images/videos if required.

[dead] 2 years ago

You should be embarrassed by the obviousness of this spam.

G1 AR Glasses 2 years ago

Maybe you would never misuse this technology, but if it becomes widespread and normalized then plenty of people are going to be walking into public toilets (other previously private spaces) with this tech on. We're creating the conditions for total surveillance of the physical world.

Is this the future you want to live in? It's not too late to back down.

EDIT: As pointed out by those below, these glasses don't have cameras. Which I think is much more preferable to the alternative.

AI App Graveyard 3 years ago

These companies are not designed to last forever. Many wrappers can make decent amounts of money for their creators for months and then die. Under these conditions a company can close down and still be considered a big success.

This seems like an Innovator's Dilemma (adjacent) trap. The actors will prevent hollywood or the mainstream studios from using synthetic actors, writers, etc. so it will be left up to some completely new entity to create that infrastructure. It will happen anyway because it is so compelling, and there are tens of thousands of actors who never make a living who would happily become digital.

* The Innovator's Dilemma - https://en.wikipedia.org/wiki/The_Innovator%27s_Dilemma

I didn't think it was that vague. But if you're looking for ideas.

1. High quality LLM made free for commercial use.

2. LLM is used in many places as it is the best.

3. LLM is aligned to subtly promote the interests of the threat actor.

There is no 'compromise'. It is not hacking software, only wetware.

A concrete example...

LLM created by Advanced Persistent Threat (APT) is used in educational software aimed at kids. Over time as the kid interacts with it the LLM promotes a way of thought that either aligns with the APT or undermines the ideology of the society they are in. There is no moment that can be pointed to that says: "look they are trying to hack us!", but decades later you have adult members of a foreign society more open to your way of thinking.

i.e. Soft Power.

There is a great opportunity for totalitarian and authoritarian regimes (China and UAE so far) to create commercially usable and free LLMs that work significantly better than alternatives (backed by large amounts of government money).

Over time, as they get used in more and more products, these LLMs can become more 'aligned' to these regimes way of thinking.

There are no Chinese companies that are not part of the Chinese government.

This is a new kind of cultural soft power.

But doesn't an engineer's authority come their skill-set and not from peoples perception of them or their job title?

When you say diminished, are you are talking from a societal point of view?

Prompt Engineering is a fun title in a fast moving and interesting field. Everyone's taking it far too seriously.

I'm going off to become a Clay Engineer, that sounds fun!

Bluesky - https://blueskyweb.xyz/blog/6-02-2023-beta-update

" Roadmap and Growth

A few weeks ago, we intentionally slowed our invite roll-out while we built more moderation tooling and capacity for users on the app. We staffed a content moderation team with shifts that cover a 24/7 schedule, and consulted with trust and safety experts to establish new processes and policies to support a growing userbase. We’ve resumed sending out daily invites to the waitlist, which is where the majority of users already on Bluesky received their invites. For those who don’t know someone personally with an invite code, the waitlist is the fastest way to receive a code, though please be patient as we work through the list. "

You may not ever be able to persuade a bot, but you can identify if it is capable of repudiating your points or not. You can ascertain if it is arguing in good faith and chose to end an argument if it isn’t.

When you say “find an argument that works” isn’t this just saying that they’ve said something you find persuasive?

What’s wrong with this? How is this different to any other argument in good faith?