Chaz
cdarv.bsky.social
Chaz
@cdarv.bsky.social
Data Dude, Culinary Entusiast
Reposted by Chaz
October 31, 2025 at 6:50 PM
He wants us to cover this stuff. This is intentional flooding the zone.

Please clear your timeliness of his words and focus on tangible actions.

His admin just had a Thursday night massacre that is under reported because of distractions like this.
American presidents don’t even joke about this. Just because he’s petty doesn’t mean this isn’t dangerous.
February 20, 2025 at 2:56 PM
We need to start ignoring what he says. Like yesterday. This is how he controls the news cycle. Treat the pres like you with trolls here: block and move on.

We should only be focusing on actions. Actions can't be waved away as easily. They have lasting impact. Focus on the FAA cuts/freezes.
I love that a white guy was flying the plane, a white guy was flying the helicopter, and a white guy gutted the FAA last week but somehow the blame is too much diversity.
January 31, 2025 at 10:24 AM
Reposted by Chaz
Reminder: Don’t. Use. ChatGPT. As. A. Calculator. Seriously.

Loved listening to @sashamtl.bsky.social on Hard Fork with @kevinroose.com and @caseynewton.bsky.social—it really made me think. 🧵
January 20, 2025 at 7:25 PM
Did anyone really want this?

Trump and his followers are such a joke
January 21, 2025 at 10:22 AM
They are laughing all the way to every corner of power in the country, unfortunately.
January 20, 2025 at 9:25 PM
Reposted by Chaz
Thank you to Redhawk Coffee for supporting the homeless homies and our neighbors leaving the jail! @man-eternally.bsky.social
January 18, 2025 at 5:02 PM
Reposted by Chaz
Tbh most of this is probably just the specifics of OpenAI RLHF "house style" but the dominant feature of LLM generated text is extremely low informational content per word and that really annoys me
January 5, 2025 at 8:32 AM
One of my favorite feeds! Just waiting for my turn to be added to the contributor list!
Welcome to What's Cooking! 🍽️

1️⃣ save/pin the feed if you're enjoying it
2️⃣ repost to help us find Bsky's next Top Chef.
3️⃣ respond with #addmepleasechef to get on the contributor list
4️⃣ use 🍽️ to post to feed

FAQ 👉🏻 bit.ly/whatscookingFAQ
January 2, 2025 at 6:18 PM
I think there's a more boring answer. Context lengths dont do as much as people like to think.

A good RAG + LLM better recall than one with a huge context length for understanding a subject.

Having it create an outline and build in parts is the way to do it rather than spit it out in one response.
Something I don't understand is: why can't LLMs write novel-length fiction yet?

They've got the context length for it. And new models seem capable of the multi-hop reasoning required for plot. So why hasn't anyone demoed a model that can write long interesting stories?

I do have a theory ... +
January 2, 2025 at 11:29 AM
🍽 spinach and ricotta stuffed chicken with steamed brocolli
January 1, 2025 at 11:58 PM
In tech I'll be saying "promoting code to open source"

I love it.
I have a colleague (not on social media) who likes to say that works don't *fall* into the public domain, they *ascend* into the public domain. I really love that.
January 1, 2025 at 1:36 PM
Happy New Year, everyone!

Let's get after it, 2025
January 1, 2025 at 6:50 AM
Reposted by Chaz
The Biggest #AI Flops of 2024

From #chatbots dishing out illegal advice to dodgy #GenAI search results, take a look back over the year’s biggest AI failures.

www.technologyreview.com/2024/12/31/1...
#ArtificialIntelligence #regulations #intellectualproperty #technology #health #democracy #humanity
The biggest AI flops of 2024
From chatbots dishing out illegal advice to dodgy AI-generated search results, take a look back over the year’s biggest AI failures.
www.technologyreview.com
December 31, 2024 at 10:14 PM
I often explain to people that are new in tech the level of care that went into mitigating damage from #y2k.

By the way, there's a funny little issue called the 2038 problem that we have to worry about soon. Maybe we'll keep the lessons we learned
Sigh. It wasn’t a nonevent BUT we took it seriously. It was a nonevent BECAUSE we took it seriously. Same dangerous attitude I’ve seen about ozone layer deterioration: if a major campaign of concerted action averted the problem, it must have been a fake problem. Not the correct lesson to draw!
Yes, the Y2K software problem turned out to be a nonevent. But 25 years ago, The New York Times took it seriously, publishing over 100 stories.

Our tech reporter who wrote a front-page article on Jan. 1, 2000, reflects on the crisis that never was.
December 31, 2024 at 10:32 PM
" #AI is useless"

Use it. Ask bing copilot or gpt to help craft a recipe with what's in your fridge.

If you want to add something, ask it follow ups. Dietary restriction? Have it help work around it.

Try huggingface chat to see open-source in action if you hate openai and microsoft

#everydayAI
December 28, 2024 at 1:16 PM
This. Or delete all your posts, unfollow, lock down your account and leave it like a graveyard of empty @'s
why the fuck would you leave twitter by deleting your account? go out with a bang, get banned.
December 28, 2024 at 12:22 AM
Diving into voice generation tonight a bit. I think I can make some goofy characters to "call" My kids with personalized messages. The open source tools developed so far are incredible.

#TTS #opensource #AI
December 27, 2024 at 11:58 PM
You should learn about #AI.

Especially if you hate it. It seems like a trend because companies are slapping it on everything, but in some form or another, generative technology is here to stay and will affect your industry.

Don't be caught flat footed because you don't want to understand it.
December 24, 2024 at 11:17 AM
Finally got around to deleting all my posts from the other site. Locked down account and removed all my followers and following.

Feels good to shed it, feels better seeing the vibes here every day
December 24, 2024 at 1:03 AM
It's amazing what you can do even with a laptop now.

#Ollama and #OpenWebUI have really made having a private local #LLM not just possible, but easy to get up and running.

The open source community around generative models is truly incredible and kinda slept on

#AI #ML
December 23, 2024 at 2:20 AM
Lot of posts about libraries in my feed.

Have to say.. I love it. Take your kids, get them a card, teach them about borrowing books.

Use their services, online books, meeting spaces.

It's one of the most important institutions we have.
December 22, 2024 at 11:17 PM
Reposted by Chaz
I don't know that I've ever heard a better idea than bagels.
December 22, 2024 at 4:36 PM
I can't second the hoopla app enough. It's a crazy resource that you get just for registering at your local library.

Most of the time you can do this all online.
Libraries are so cute. Please get the Libby and Hoopla app on your phone and the Kanopy app for your TV. It makes the libraries so happy and you can access books, movies, magazines, audiobooks, tv shows, and joy. YOU GET FREE JOY
December 18, 2024 at 11:56 PM
It's beans and greens weather
December 15, 2024 at 6:33 PM