Tim Kellogg
banner
timkellogg.me
Tim Kellogg
@timkellogg.me
AI Architect | North Carolina | AI/ML, IoT, science

WARNING: I talk about kids sometimes
Pinned
Does AI get bored?

I gave them nothing to do, just to see what happens

one thing — they devolve into a repetitive “collapse” state, I guess you could call it boredom

but some break out into math & poetry on their own, I didn’t expect which ones that would be

timkellogg.me/blog/2025/09...
Does AI Get Bored?
timkellogg.me
but is it still learning if you got it from an LLM?
November 10, 2025 at 12:34 PM
sent this to my brother asking, “does this count as wealth redistribution?”

(fun fact: my bro voted for Trump and is also undergoing collapse of the company he’s CEO of due to tariffs)
November 9, 2025 at 9:36 PM
ok now i’m just curious
sex workers who make their money on the internet often hang out being cool people on the internet to generate traffic on a sfw account.

if you have zero sex workers as mutuals it means you're using the platform you're on like linkedin and checking their resumes in advance
one comment on the Talarico Instagram non scandal: sex workers are allowed to have lives and interests outside of work, too, including politics.

feel like bsky is preaching to the choir in that regard, but
November 9, 2025 at 9:09 PM
The town of German, NY elected 2 positions on write-in ballots alone

1. Superintendent of Highways
2. Town Justice

apparently no one ran
November 9, 2025 at 9:06 PM
social media is RL on humans
It's interesting that RLHF'd LLMs and influencers talk the same way. Perhaps through the evolution of clickbait, we'd already found the local maximum of attention grabbing
November 9, 2025 at 8:24 PM
most engineers probably don’t understand the extent to which “tech debt” works as an abstraction

there’s absolutely “good tech debt”, and it’s not always obvious which you’re dealing with

if you’re green fielding an app, you should expect to launch with some amount of tech debt
November 9, 2025 at 5:20 PM
why has nobody bothered to ask the hard questions that matter??

like, how can we appear to address income inequality while actually shifting even more wealth to the ultra rich?
November 9, 2025 at 2:47 PM
this is true
The pope: “you should probably be a good person”
Marc Andreessen: “this is an attack on me and everything I stand for”
November 9, 2025 at 2:25 PM
Polaris Alpha, believed to be GPT-5.1 non-reasoning, scores just below Sonnet 4.5 on HLE (unofficial run)

There will be a reasoning version too, and OpenAI excels at RL & post training, so I have high expectations for it

also leaked: Nov 24 release date
November 9, 2025 at 2:18 PM
idk is a 50 year mortgage even worth it?
November 8, 2025 at 10:45 PM
pro tip
November 8, 2025 at 10:20 PM
kimi-writer: an agent that writes long-form fiction using K2-Thinking

it writes files using tools, auto-compacts context when it grows too big

seems like the convergence of high-quality writing and agency

github.com/Doriandarko/...
GitHub - Doriandarko/kimi-writer: AI writing agent powered by kimi-k2-thinking - autonomously creates novels and stories with deep reasoning
AI writing agent powered by kimi-k2-thinking - autonomously creates novels and stories with deep reasoning - Doriandarko/kimi-writer
github.com
November 8, 2025 at 10:01 PM
Reposted by Tim Kellogg
An increasingly likely future is that China reaps the benefits of the world's most advanced AI systems

Trained on Chinese chips

Using free Chinese renewable energy

Because the American right turned it's back on low-cost clean energy and free trade

And the American left turned it's back on AI
Did you know a former White House climate advisor flipped a GOP stronghold in Virginia by running entirely on putting a stop to more AI data centers?

For @heatmap.news I profiled John McAuliff and a campaign that will be a roadmap for all future anti-AI politicians moving forward.
This Virginia Election Was a Warning for Data Centers
John McAuliff ran his campaign almost entirely on data centers — and won.
heatmap.news
November 8, 2025 at 3:46 PM
GPT-5-codex-mini

Almost same performance as GPT-5-codex on high, but 4x faster and without pesky things like warm personality

www.neowin.net/amp/openai-i...
November 8, 2025 at 4:46 PM
from talking to my healthcare friends, one ironic thing they’ll do is take all their structured data and cram it into an unstructured blob of text to do analytics on it

this was a thing even before LLMs
November 8, 2025 at 4:40 PM
“nah, we don’t do 996”
November 8, 2025 at 12:49 PM
this morning, X is saturated with people from US claiming that their favorite unknown benchmark (that happens to show K2 trailing US models) is actually the best single benchmark to watch

lol notice how they clipped off the top 12
November 8, 2025 at 12:10 PM
November 8, 2025 at 11:01 AM
i feel like every “unsupervised” training algorithm starts to look more supervised as you sink endless hours into it
November 8, 2025 at 2:15 AM
great post!
When I started my doctorate, I had to complete the standard research ethics training. Particularly with prisoner populations, I saw many parallels to AI. #ai #artificialintelligence #llms #largelanguagemodels #aiethics
When We Decide Who Can Feel
Should AI be protected by ethical research guidelines?
open.substack.com
November 8, 2025 at 1:55 AM
oh wow, any of this stuff would be incredible, but the steering vector stuff in vLLM, yes please
@vgel.me is fundraising for her model tinkering, she's done some really interesting interpretability work and I think funding this has very high returns in terms of LLM understanding per dollar. manifund.org/projects/fun...
November 8, 2025 at 12:03 AM
wow, i had no idea
November 7, 2025 at 8:26 PM
K2-Thinking is available in the Kimi app now
November 7, 2025 at 7:29 PM
longer form position here

www.vaticannews.va/en/pope/news...

i really like this part
November 7, 2025 at 6:35 PM
i hate how much work “a paper” is doing here

like, i could probably send a coworker an arXiv link about LLMs ability to write COBOL with the tagline, “you think we should switch?” and they’d probably take me seriously

it’s getting ridiculous. The Economist should know better
November 7, 2025 at 5:34 PM