Joined December 2008
- Tweets 33,266
- Following 4,624
- Followers 2,681
- Likes 38,060
3,147 Photos and videos
Pinned Tweet
6 May 2017
Now available on YouTube: My #TEDxASU talk about the hidden costs that disciplinary identity puts on innovation.
youtu.be/9GWXCRetOjk
2
11
54
Colleagues are often curious about how best to integrate LLM's into #Rstats workflow. This guide covers different options from direct integration into #R IDE to keeping #LLM separate. Should apply to #ClaudeCode, #Codex/#ChatGPT, #Gemini CLI, #Copilot, etc
github.com/tpavlic/docs-and-…
30
So, let me get this straight. Investors thought other investors would confuse SPCE for SPCX, and so they INTENTIONALLY bought SPCE in anticipation of that, leading to a rise in SPCE stock...
1
95
Ted Pavlic (he/him/his) retweeted
Jun 11
I've been telling people for 25 years that Jane Street is not interested in formal methods.
No more!
And we're actively hiring to form a new formal methods team!
29
72
1,067
119,912
Might have been more interesting to discuss how disciplinary partitioning (math vs CS) sometimes means topics like GIT emerge in one but live out lives in another, hidden from former. Like Russell's barber, we've created a set of all disciplines that is itself not a discipline.
Jun 10
Gödel's incompleteness theorem is one of those things that math popularisers always talk about as a hugely important result.
But according to @3blue1brown, it almost never comes up in practice. It's a weird pathology that nobody expects to matter for the big questions.
84
Interesting approach: "We’ll teach 1,000 fellows how to use Claude well, match them with nonprofits across America, and pay them to spend a year—full-time, in-person—helping host organizations to advance their missions."
"Introducing Claude Corps"
anthropic.com/news/claude-co…
1
40
Lately, I've been thinking a lot about multiscroll hyperchaos and different application spaces than the conventional image encryption case. To help discuss these the topic in general, I created this new primer (which doubles as a review of what is chaos).
tpavlic.github.io/topic_visu…
24
Ted Pavlic (he/him/his) retweeted
Jun 10
wellWell
11
32
315
27,524
Ted Pavlic (he/him/his) retweeted
Now updated to include a third tab that covers using mixed-effects Cox Proportional Hazards models for hypothesis testing (and various visualization options).
tpavlic.github.io/topic_visu…
1
58
To help a student understand how to handle right-censored data, I built this survival (and recurrent-event) analysis explorer that demonstrates the approach and provides some basic derivations. Applications to compartmental models from epidemiology too.
tpavlic.github.io/topic_visu…
1
59
Now updated to include a third tab that covers using mixed-effects Cox Proportional Hazards models for hypothesis testing (and various visualization options).
tpavlic.github.io/topic_visu…
1
58
Do you think @AnthropicAI called it "Fable" because there are just unverified tales of it actually being willing to complete a request?
#Fable5
90
I don't know how, but somehow Claude has spontaneously decided to check its own responses for em dashes and make sure to construct responses to me without them. I guess I yelled at it enough that it started remembering?
2
71
Ted Pavlic (he/him/his) retweeted
Jun 7
Prepare yourselves for the outcry when the demand for University education continues to rise and becomes even more expensive in the LLM era
Hot take: Universities charge $300,000 for a degree that teaches you skills any LLM can do for free. At some point we need to have an honest conversation about whether higher education is the greatest individuals misallocation of capital in recent history.
24
8
119
14,470
Ted Pavlic (he/him/his) retweeted
Jun 7
just overheard a guy introducing himself as a “backend software engineer” and a girl reply “backend? omg happy pride month” LMAO
55
1,913
46,701
911,583
From UI to UX to... "AX"? -- "According to Lawson, [engineers] are the ones who understand what agent experience (AX) can and should be."
"Netlify CTO Dana Lawson: Writing code is no longer the job"
thenewstack.io/netlify-agent…
30
So glad we have so many dancing #robots...
#WhereIsMyJetpack
43
Imagine if every article had to include, "This manuscript was assembled using a word processor. In some cases, a calculator was used. Handwritten work was done on a tablet which sometimes involved virtual graph paper and a virtual ruler. Plots were generted with a graphing..."
Jun 5
If Claude is good enough for Nobel Prize winners it is good enough for you
arxiv.org/abs/2606.03300
2
1
6
611
It's not like these statements help with reproducibility nor absolve the authors of responsibity for research integrity violations. I guess they could ask for chat transcripts, but what if they were lost or extremely long and not timestamped and so impossible to interleave.
54
A new product uses lidar to find mosquitoes and zaps them with a laser. Oh, what a world...
photonmatrixlab.com/
40