kellogh

kellogh, 3 months ago to climate

this is very encouraging. Fuel cells have always been a wild fantasy that doesn’t quite work in the real world, but it seems like a few forces in politics and industry are changing that

https://www.axios.com/2024/02/07/hydrogen-vehicles-fuel-cells-emissions #climatechange #climate #hydrogen #fuelcell #cleanenergy

kellogh, 1 year ago to science

this is incredible. Hanging solar panels a few meters above crops of tomatoes and jalapeños multiplied their yield 2-3x, used substantially less water, controlled temperatures, and increased the output of the solar panels https://www.pnas.org/doi/10.1073/pnas.2301355120 #science #solar #environment #agriculture

kellogh, 1 month ago to random

The Attack Turkey

this morning on the run i got to the top of the hill and was resting, bc i'm out of shape, and off in the distance i see a bird trotting my way

as it gets closer i realize it's a turkey. it had a big red gobbler and its neck was iridescent. It was actually really pretty, so i just watched as it approached

kellogh, 7 months ago to ai

i appreciate #ai skeptics, they have valuable points. but it annoys me when they double down on some point that’s totally bogus

one example: #llms aren’t reliable and therefore we can’t use them for engineering

it’s misguided, but it’s worse than that. we engineer systems out of unreliable components all the time, that’s a core responsibility of engineers

e.g. amazon s3 has 11 9s if durability, from disks that have less than 99%

e.g. bridges & skyscrapers are built from unreliable materials

kellogh, 6 months ago to LLMs

i’ll say it — #LLMs can and will spit out any topic they’ve been trained on

an absurd amount of research is going into preventing the #LLM from explaining how to make a bomb, when they could just do some dumb tricks and remove the “how to make a bomb” manuals from the training corpus.

am i missing something?

kellogh, 4 months ago (edited 4 months ago) to random

Last night I came up with (and implemented!) an idea for a mastodon client that automatically curates my feed by categorizing toots. It's just idea phase right now, but I wrote about the process here https://timkellogg.me/blog/2023/12/19/fossil #LLMs #AI #feditips

kellogh, 7 months ago to random

dear internet, where tf did Pascal_Snake_Case come from? is there a name for it? is it a real thing people try to do or is it only used by people badly mimicking other naming schemes?

kellogh, 4 months ago (edited 4 months ago) to Minecraft

what’s an appropriate age to start letting #kids play #minecraft?

kellogh, 3 months ago to random

imo XML gets a lot of hate because it's for semi-structured data but people used it for structured data.

if you have text and you want to imbue more meaning on top of it, XML kicks serious ass

kellogh, 11 months ago to random

Took a paddle on the mystical Merchant’s Millpond today. Surreal. Like paddling through a dream

kellogh, 3 months ago to random

i’ve become disappointed with the lobste.rs community lately. it used to be the best technical discussion on the internet, to the point that i would skip the article just to read the comments, they were that good.

but the comments on my latest blog were 100% about packaging stuff, like appearance, linking, etc. it’s weird but hacker news is now where the real conversation happens and lobste.rs became a cesspool

https://lobste.rs/s/xnzvea/htmx_is_composable

kellogh, 1 year ago to random

I hooked the ChatGPT API to speech recognition and now the kids can use it. They’re thrilled, although annoyed that it refuses to tell them fart jokes

kellogh, 7 months ago to LLMs

one of the “business person” talking points on #LLMs that annoys me is “memories”. they’re wowed at an AI’s ability to remember things, as if 2 TB hard drives didn’t exist

kellogh, 9 months ago to python

hot take: dynamic languages like #python and #javascript would benefit a lot by a little type checking around async/await constructs — simply emitting a compiler error (or even a warning) when you don’t await an async function call would save massive amounts of time

kellogh, 10 months ago (edited 10 months ago) to random

if you bake 2 cakes and stack them on top of each other, how many cakes do you have?

kellogh, 6 months ago to LLMs

i wish i knew more about comparing #embeddings. anyone have resources? one thing i’ve wondered is how to convert an embedding from a “point” to an “area” or “volume”. e.g. an embedding of a 5 paragraph essay will occupy a single point in embedding space, but if you broke it down (e.g. by paragraph), there would be several points and the whole would presumably be at the center. is there a way to trace the full space a text occupies in #embedding space? #LLMs #LLM #AI #NLP

kellogh, 5 months ago to random

uncomfortable fact — the energy cost of irrigated corn is almost twice that of rain-fed corn. and irrigation is needed more as lands become unexpectedly dry due to climate change

kellogh, 9 months ago to random

mind control is not real

If someone says "Fox News caused this", they're invoking mind control, which is not real. Fox had to work within an existing system of values to create derived ideas. We can also work within that system of values.

Lately, I've noticed the dialog has become "climate change is destroying the world, yes, but it's fine because it's God's plan.

kellogh, 4 months ago to random

periodic reminder that the NYT v OpenAI lawsuit is just billionaire hunger games. NYT is trying to open a new revenue stream via lawsuits.

Even if GPT was reproducing the articles (it's not, in a meaningful way), those would be 3-year old articles. NYT is a news agency, nobody was going to buy 3yo articles anyway. There's no lost revenue.

totally fine if you want to pick sides, just realize that you're not fighting for the little guy

kellogh, 3 months ago to random

i have a hard time believing that using terms like “artificial intelligence” are causing people to believe that it’s “thinking”.

you don’t have to explain to people why artificial grass doesn’t grow. the “artificial” part says that on its own

“lab grown” diamonds are a thing now, they avoided “artificial” diamonds because they’re chemically identical to the real thing

besides, there’s far simpler explanations — e.g. the “AI is going to take over the world” messaging

kellogh, 11 months ago to random

imo the user experience of GitHub #Copilot stinks. Generating code is one of the tasks I trust an #LLM least. I’d rather a chat interface so I can ask it to

1. Refactor
2. Generate files
3. Move files
4. Navigate
5. Ask questions about code
6. Understand a new code base

Sure, writing little bits of code is kinda cool, but also ehh 🤨 I’d rather just type it myself. Feels like a lot of opportunity left on the table

kellogh, 3 months ago to 3DPrinting

#3dprinting has been a lot of fun, but i don’t see it scaling out to general audiences. simple things like printing an existing model are pretty complicated. even just, “load model, switch spool, print” is far beyond what my 7yo can do, and that seems like a big UX problem

i wonder if #LLMs could help parts of the UX. load a model and the LLM asks what you’ll be using it for, adjusts infill & speed parameters appropriately. idk, the whole market seems dead without something big changing

kellogh, 1 month ago to random

I would love to see a non-profit that's basically, "Apache for contributors". Apache is great for companies who want to de-risk open source software. What about an org that goes, "i see that your project is important to the community, what can i do for you today?" Maybe fund, maybe find contributors, maybe acquire build infra, always listen and support

kellogh, 1 year ago to random

Dads against the Routine #metal #incendiary #hardcore

kellogh, 2 months ago to random

why would you want a machine to be conscious? setting aside the debate of if it is, what would that let it do that an unconscious machine couldn't?