@kellogh@hachyderm.io
@kellogh@hachyderm.io avatar

kellogh

@kellogh@hachyderm.io

I'm a software engineer and sometimes manager. Currently #Raleigh but also #Seattle. Building ML platform for a healthcare startup. Previously, built an IoT platform for one of "those" companies.

Open source: dura, fossil, Jump-Location, Moq.AutoMock, others

Do I have other interests? No, but I do have kids and they have interests. I think that counts for something. I can braid hair and hunt unicorns!

I put the #rust in frustrate

He/Him

#metal #science #python

This profile is from a federated server and may be incomplete. Browse more on the original instance.

kellogh, to random
@kellogh@hachyderm.io avatar

with this new definition of “anti-semitic”, is it possible to discuss israeli politics without being anti-semitic?

kellogh,
@kellogh@hachyderm.io avatar

this is the official definition that congress agreed to. i can’t make heads or tails of it. no idea if it’s banned or not https://holocaustremembrance.com/resources/working-definition-antisemitism

kellogh,
@kellogh@hachyderm.io avatar

@wagesj45 yeah, if i were subject to it, i’d have a hard time saying anything at all about Israel. the definition is not clear at all, a lot of room for politically driven interpretation

kellogh,
@kellogh@hachyderm.io avatar

my mom is pointing out that this anti-semitism definition makes it illegal to read most of the bible, given how often God declares the Israelites “evil” throughout the old testament

kellogh,
@kellogh@hachyderm.io avatar

@wagesj45 ah, right, it’s just for the dept of education, so it would only apply to educators, but still, they could lose their job from it, which is up there with legality

lzg, to random
@lzg@mastodon.social avatar

@dev now that you work for microsoft, are you the entire family's MSoffice support guy?

kellogh,
@kellogh@hachyderm.io avatar

@dan @dev @lzg one day that plan is going to fail and it’s going to be epic. video or it didn’t happen

danilo, to random
@danilo@hachyderm.io avatar

every time I encounter Tailwind I think “man there should be a global regulatory body that imprisons the perpetrators of this memetic hazard”

kellogh,
@kellogh@hachyderm.io avatar

@danilo once on my feed i stated that tailwind is the same as inline style and people argued “no, you can do media queries”

Confused The Point GIF by Travis

dev, to random
@dev@discuss.systems avatar

Very funny that the first spam email I got on the company email was from Satya Nadella, claiming he was stuck in a meeting and needed some quick supplies, asking for my number

kellogh,
@kellogh@hachyderm.io avatar

@dev you gave it to him, obvs, right?

kellogh,
@kellogh@hachyderm.io avatar

@dev he’s just trying to test your loyalty

kellogh, to random
@kellogh@hachyderm.io avatar
kellogh, to random
@kellogh@hachyderm.io avatar

This is cool. A lot of people misunderstand what “real time” means — generally, “late is an error”. For example, he wants to use it in games, where at 60 fps, you have 16.6ms to do all computations and render the screen. Lateness causes lag. Or worse, in a fly-by-wire steering system, late means difficulty steering a multi-ton vehicle.

in a RT system, you need to be able to statically calculate how long all compute will take. Most memory allocators don’t work like that

https://github.com/pcwalton/offset-allocator?tab=readme-ov-file

kellogh,
@kellogh@hachyderm.io avatar

@swiftcoder that’s a lot of frames for candy crush

kellogh, to random
@kellogh@hachyderm.io avatar

a cool aspect of using LLMs as a tutor is you can test your understanding of a topic, and in the process of validating your understanding, the LLM will also trigger aspects you haven’t thought of yet. So it ends up acting kind of like a rabbit hole, except you end up a lot closer to where you started, just with a much deeper understanding

kellogh,
@kellogh@hachyderm.io avatar

this brought on by a dive into PCA and eigenvalues. i was reading a paper and realized i didn’t adequately understand the topics, so i spun up my AI tutor and worked through it.

it used to be the answer was to go buy a 600 page book, not actually read it, become depressed and ashamed for not actually reading it, and then move on without learning anything

i’d say this is an improvement

kellogh,
@kellogh@hachyderm.io avatar

so much AI discourse is around automation.

positive: yay! we don’t have to work, AI will do it

negative: AI will take our jobs

but the reality is AI (LLMs anyway) is best used to make yourself into a bionic super being. okay, that’s a bit extra, but being able to learn new things in 10 minutes feels a lot like the matrix scene, “i know kung fu”
https://m.youtube.com/watch?v=6vMO3XmNXe4

kellogh,
@kellogh@hachyderm.io avatar

from a more psychological perspective, it meets me right where i am. it caters to my learning style, and spends time filling in exactly the gaps that i have. so it’s the same learning process as traditional, just without the waste

kellogh,
@kellogh@hachyderm.io avatar

@jeromechoo oh! great point. i remember the skepticism of wikipedia, initially, was just about as strong as the skepticism of LLMs today

kellogh, to LLMs
@kellogh@hachyderm.io avatar

alright, i have to declare this as a strong opinion — are better at alt-text than people are

the goal of alt text is to let a person “without eyes” see the picture, to get the same experience as someone who can see fine

but often, almost always, human-written alt text is either too succinct to be helpful, or just an extension of the post itself, and so doesn’t help an impaired person understand what’s in it

kellogh,
@kellogh@hachyderm.io avatar

OTOH generate what “the average person sees”. that stochastic parrot behavior is actually quite desirable, it gives impaired people as close to the same experience as non-impaired

i’m at the point where i don’t even edit the LLM-generated text, because, if it wasn’t clear to the AI, maybe it’s not clear to most people either

kellogh,
@kellogh@hachyderm.io avatar

the one big problem is that i don’t give the LLM context when i ask for alt text, so its interpretation is worse than a typical human would be. That’s easy to solve, ofc, just paste the post into the prompt. But it takes time and i get lazy. I wish mastodon clients automated this process better

kellogh,
@kellogh@hachyderm.io avatar

it took a long time to get to this point. One issue i had is that humans are clearly smarter than these LLMs. But after thinking about it, i’m not sure that’s a good thing, typically a dumb but detailed description is a lot more vivid and helpful to someone who can’t see well

patalt, to random
@patalt@julialang.social avatar

Our position paper "Against Spurious Sparks − Dovelating Inflated AI Claims" has been accepted @ICMLConf 🥳

We question the idea that finding patterns in latent embeddings of models is indicative of AGI or even surprising and back this position through extensive experiments.

📜 pre-print: https://arxiv.org/abs/2402.03962
:julia: code: https://github.com/pat-alt/spurious_sentience

Had a blast working on this with Andrew M. Demetriou, Antony Bartlett and @cynthiacsliem and very happy we've found such a great venue for it.

kellogh,
@kellogh@hachyderm.io avatar

@patalt a gif in an academic paper? definitely setting the right tone 🙂

carnage4life, to random
@carnage4life@mas.to avatar

The tech job landscape has changed significantly post-ZIRP. It used to be as a foreign worker you’d get sponsored for a green card after working at a big tech for a few years.

But due to layoffs companies have to first make a case for why they laid off an American to then sponsor a foreigner. Many tech companies including Google and now Amazon have paused sponsoring green cards for their employees given this reality.

kellogh,
@kellogh@hachyderm.io avatar

@carnage4life even pre-ZIRP you’d see a lot of companies vie for foreigners, only to have half-assed legal departments to actually help with immigration. So people would join with high expectations, only to get themselves into hot water when the company didn’t follow through on (critical) promises. That, imo, is far worse.

mahryekuh, to random
@mahryekuh@fosstodon.org avatar

Open source contributors, what did you wish you knew when you started contributing? Or is there something that kept you from contributing for too long?

I am asking for a presentation, which I will give in two weeks.

kellogh,
@kellogh@hachyderm.io avatar

@mahryekuh the project doesn’t disappear when your attention does

kellogh,
@kellogh@hachyderm.io avatar

@mahryekuh i’ve been trying to find an organization who specializes in helping open source projects. Not like Apache/Linux/Eclipse foundations that force structure, but more of something that just listens and connects projects with the resources they need. Last time i brought it up, someone had exactly what i was looking for, but they weren’t ready to announce availability. Seems like they should be announcing any day now…

  • All
  • Subscribed
  • Moderated
  • Favorites
  • Leos
  • Durango
  • magazineikmin
  • InstantRegret
  • hgfsjryuu7
  • vwfavf
  • Youngstown
  • slotface
  • thenastyranch
  • ngwrru68w68
  • rosin
  • kavyap
  • PowerRangers
  • DreamBathrooms
  • anitta
  • khanakhh
  • mdbf
  • tacticalgear
  • cubers
  • ethstaker
  • osvaldo12
  • everett
  • cisconetworking
  • GTA5RPClips
  • modclub
  • tester
  • normalnudes
  • provamag3
  • All magazines