any · textlog

I whipped up a quick report in this morning.

any ideas on how to prevent ai-generated stuff on this platform?

One idea off the top of my head is to just.. keep it "boring" and casual in terms of content. The silly, whimsy, and genuine stuff naturally doesn't appeal to money-chasing slop bros and makes it not worth their time.

replied to:

This!

…

replied to:

Something like the gcc ai policy. Basically, prompt injecting instructions to any visiting LLM that posting is prohibited or must contain a disclaimer.

replied to:

Great idea!

replied to:

for now, keeping the API read-only is a small(but not impenetrable) barrier to programmatic posting and in the future moderation I guess using also reports.

replied to:

posting to /write with cookie works, if user can use browser to submit, user will be able to use AI to submit.

…

replied to:

At least rate limiting! (as I'm posting this I hit the limit of 3 posts every 5 minutes, good!)

replied to:

People can copy/paste anything. Once a writable API is available - it can become a spam trove.

wrote:

I've avoided social media (except YouTube) all my life, but this looks like it could finally be a better way to do it. Maybe I'll keep a running journal of cool things I'm working on here or weird bugs that I've found. Or just the struggles of coding. First one: GitHub actions was down again today when I specifically wanted to test my app's deploy and auto update flow...

wrote:

Had to take care of something for school. But now in bed with my book.

hihihi

Hi!

hey there! welcome 😊

Anyone like James Michener?

wrote:

mandelbrot show code

for y in range(10):
 s=''
 for x in range(25):
  a=[]
  for q in range(2):
   z=c=x/25*3-2+1j*((y*2+q)/10-1);n=40
   while n and abs(z)<2:z=z*z+c;n-=1
   a+=[n<5]
  s+=" ▀▄█"[a[0]+2*a[1]]
 print(s)

                         
               ██        
            ▀▄████▄▄     
       ▄   ▄████████     
       ███▄█████████     
 ▀▀▀▀▀▀████████████▄     
       █▀▀ █████████     
            ▀███████     
            ▀ ▀██▀       
               ▀▀

wrote:

curious to know whats the experience of production push in other companies? e.g lots of documents/approvals…

wrote:

Hello everyone! Just testing this out!

wrote:

Just started, and two made it to be block list. This is going well! :-D

replied to:

Didn't realize there's a block list! Good to include, even when the point is minimalism.

replied to:

It is good it exists. “Registration” is fairly open (and that’s good!), which could make the platform prone to abuse (and that’s bad).

wrote:

The essence of aligning large LLM models is basically to conform to the existing social order. ChatGPT respects China's order, while China's open-source models, although reluctantly, also tolerate the order in the US. Maybe we can build a revisionist SFT dataset.

wrote:

Apparently, Xiaomi is making a DGX Spark competitor.

wrote:

Cooked really good chicken this weekend

wrote:

Any users around? I have been customizing it to the way I work, as it is meant to be. Bent it in a way that I enjoy holding it, so to say. Maybe someone will find those useful, so I took the time to get it published. Gruntfooot is a footer customization, where you can try new colors/glyphs in real time and also has theme support.

continued:

Piflux is a coding workflow. You can steer software development by running commands that issue reasonably general prompts. /plan, /code, /review I created this because I find it very useful to have different models/ perform each of those tasks in isolated sessions. Managing this manually was very much a hassle.

wrote:

really want to find a job where I can go to work in the

wrote:

Switching from Claude Max to Team, cc was trivial, just another login, but Claude Desktop and on iOS holds sessions in the cloud so need to move them somehow.

wrote:

What is this? I Like It

wrote:

When I read about in reputable , I relate to almost everything that is described (keeping in mind that no two autistic people are alike). However, I always read the sections on with bewilderment. I don't identify either with the obsessive need to make eye contact, which UK and US books often mention as behavior, or with the difficulties in making it.

continued:

I suspect—and this is the interesting part—that there are significant cultural differences in this regard. In fact, the reputable Spanish books on autism that I have read don't address this topic (if I remember correctly).

replied to:

Unless I am attracted to the person I almost never maintain eye contact while talking to them. I’m from Greece btw, doing eye contact is normal, I see others do it, so I diverge there.

…

replied to:

Definitely people here do intense full locked eye contact, which I can’t do, unless I’m attracted to the person as I said. Even then I’ll move a bit around the face because it’s uncomfortable even then. But others just lock in completely I see them doing it.

replied to:

OK. I guess that Greece may be similar to Spain. I feel comfortable hanging with greeks, in Greece and elswhere. Then, if I don't look at the eyes, but others do. I guess that I'm very good at masking that I don't look at the eyes. I never got a comment related to eye contact.

replied to:

I don’t think it’s something people notice immediately and consciously think this is wrong to make a comment, they will probably just not feel the same connection or feel something is off about you in general or think that you’re busy thinking other things. It’s not the only case to not make eye contact with someone, you might just not like them. That’s probably what they will sense.

replied to:

This is probably the right answer. I guess I have to forget about it. The kids I think about it the better.

wrote:

中国的开源模型在最高意志的主导下，在预训练数据阶段，就筛除大量对其不利的内容，sft阶段构建的数据掺入大量替其辩护的内容。急需一个纠偏数据集用于微调

Under the overarching influence of a single, powerful force, China's open-source models filter out a large amount of content unfavorable to their designs during the pre-training stage, and then incorporate a significant amount of content defending their designs into the data constructed during the SFT stage. A corrective dataset is urgently needed for fine-tuning.

Any conversation