soultrees

Run an agent on it. This is what they excel at.

Not much to add other than: I love this.

I think the OP is making a point that we should resources as The Internet Archive as services that are in the best interest of the public.

Not take from the beauty of these trees, but this time of the year is always my favourite time in Vancouver for the sheer amount of colour and pop that the city gets. Sometimes only for a few weeks only to be gone in the first storm and other times lingering for up to a month of beauty. Nature is fantastic.

I’ve always thought tobacco companies had some influence on Netflix’s programming. The scenes where smoking is involved are usually set up so the whole ritual is seen (take the pack out, maybe tap it, pull it out, light and take your first big inhale with all the sounds and smells that come with it). That’s the ritual that many smokers know and has for sure influenced my cravings when I was trying to quit smoking a few years ago.

AI ownership most likely. Preventing LLMs from using it’s code in an output and not putting it in the ‘safe to spew’ bucket.

Ah, another factorio fan I see.

They don’t save money when you have to ask it multiple times to get the expected output.

What IRC Channels do you frequent?

How can you hate a technology with a passion? Sure, EVs provide some issues today but genuinely curious how someone with any hint of intellectual curiosity can ‘hate’ a technology.

Perhaps the hate the culture or the messaging around the technology right now, but ‘hating EVs with a passion’ seems weird.

No. It’s probably the last place on the internet where respectful discussion happens in the open like it does.

Whereas I felt the opposite the of Reddit. I cut out Reddit completely during that last debacle and I would say my life is better for sure.

This is the one of the first items I load in when I start a new project as it makes it so much easier to visualize the data structure when each nested level is a different colour. Thought I would polish it up a bit and share it for those who may find it useful.

Cheers.

The game here is to get as many of your users as possible and keep them chatting with the LLM (copilot) to get the human / machine back and forth. When everything has been scraped already then the only advantage you can have is more chat logs. I suspect that we will see every large tech company start to push their own models (or skins) just as hard for that same reason. OpenAI is sitting on some of the most valuable data from their user base, and the long targeted conversations they have.

The amount of effort gone for these designs is unheard of in the generative day and age. At first I thought it was Ai generated design but the text is too clean, and the imagery way too relevant. Super impressed.

.

I’m with you on that. I’ve had a north face bag for almost 20 years (and it looks like it!) but it’s still strong as can be.

The unfortunate thing is I was in the market for a new bag and went to the north face store of course and the only bag they had in the whole store was some shoestring style bag. Quite disappointing to see north face forget their roots.

Thank you for this.

Please do.

Haha yeah, I saw that after. In that case, great job everyone!

What an amazing resource. Thank you for this. And even more thanks to whoever put this together.

The most success I’ve had for classifying purposes so far is using function calling and a hack-solution of making a new object for each data point you want to classify for the schema open AI wants. Then an inner prop that is static to place the value. Then within the description of that object is just a generic “choose from these values only: {CATEGORIES}”. Placing your value choices in all capital letters seems lock it in to the LLM that it should not deviate outside those choices.

For my purposes it seems to do quite well but at the cost of token inputs to classify single elements in a screenplay where I’m trying to identify the difference between various elements in a scene and a script. I’m sending the whole scene text with the extracted elements (which have been extracted by regex already due to the existing structure but not classed yet) and asking to classify each element based on a few categories. But then there becomes another question of accuracy.

For sentence or paragraph analysis that might look like the ugliest, and horrendous looking “{blockOfText}” = {type: object, properties: {sentimentAnalysis: {type: string, description: “only choose from {CATEGORIES}”}}. Which is unfortunately not the best looking way but it works.

Thank you, I should have been more specific. I guess what I’m asking is, how deterministic would you say it is in your experience? Can this be used for classifying purposes where the values should not be outside what’s given in a variable input prompt , or when we say deterministic are we saying that , if given the same prompt then the output would be the exact same only? Or is the seed a starting parameter that effectively corners the LLM to a specific starting point only then depending on the variable prompts, potentially give non-deterministic answers?

Perhaps I’m misunderstanding how the seed is used in this context. If you have any examples of how you use it in real world context then that would be appreciated.

Fair enough and your criticism is appreciated.

Can you explain how this works more?

I’d like a productive comment if you have it.

Here come the innovation sponges.

If this goes through then the models that the general public have access are going to be severely neutered while the ownership class will have a much better model that will never see the light of day due to legal risks and claims like this - therefore increasing the disparity between us all.

Is there any open source front ends out there? I know of anything LLM but hoping to plug my own home built rag system into a nice front end.

I just did this today using openAI’s function calling. I have a bunch of elements in a scene and trying to classify them between various ‘buckets’ has been the challenge. The way I set it up is that in the schema expected, you take that top level free text, wrap it in quotes and it becomes the parent object with the elementCategory as a string required inside it, with a list of all category types in the description. Then loop through and create a dynamic schema based on how you chunk your data, then add a validation step at the end to ensure that chatgpt doesn’t forget, add or change keys. I found that if you create a numberedKeywordString for your chunk, and wrap each in quotes, then stuff that numberedKeywordString into the system prompt then it’s solid and will ensure the LLM catches every key and gets the associated elementCategory value.

It works quite well for dynamic classification in my purposes but I’m sure there is a better way.

At this point who cares honestly. The more ‘fake’ generated nudes out there, means it’s just not going to be a novelty. And if everyone has the ability to generate an image of everyone naked, the value for ‘real’ nudes will go high but it will also be good cover for people who get their nudes leaked.

That’s beautiful. I love it.