Kostas @christidis

working notes kostas.net CA Joined July 2014

Tweets

1K
Followers

420
Following

240
Likes

78K

Kostas @christidis

2 hours ago

Stumbled onto this park while walking around downtown Los Gatos the other day. Much better hang than the town's Plaza Park. Bachman Park:

0 0 0 30 0

View Details

Kostas @christidis

2 hours ago

A masterpiece, no notes youtube.com/watch?v=2WvF1E…

0 0 0 22 0

View Details

Kostas @christidis

a day ago

@MeowserMiester @levelsio are you hallucinating

0 0 0 63 0

View Details

Kostas @christidis

a day ago

@NickADobos 🤝

0 0 0 127 0

View Details

This is Professor Roland Fryer on @triggerpod discussing the reaction to his paper, particularly its findings on police use of lethal force. He received death threats, required police protection (ironic), and was suspended from the Harvard faculty over a handful of jokes that supposedly created a “hostile work environment.” His research lab was shut down for good measure. The old Twitter before @elonmusk bought it, flagged and suspended accounts simply for posting the paper, branding it as hate speech. They had endless tolerance for Arday’s lies. They went after Fryer for speaking the truth.

Melissa Chen @MsMelChen

2 days ago

To understand just how irredeemably messed up academia is, compare the differences in how Cambridge treated Jason Arday and how Harvard treated Roland Fryer. One is a black sociologist; the other is a black economist. At this point, we are all very well acquainted with Jason

168 2K 10K 610K 2K

84 1K 7K 331K 2K

View Details

Kostas @christidis

2 days ago

@Stammy To make matters worse: the default card is at the button of the stack, which never made sense to me.

0 0 1 198 0

View Details

Kostas @christidis

2 days ago

This is a wonderful shade of blue. Hats off to whoever was driving a 911 in it the other day. media.porsche.com/paint-to-sampl…

0 0 2 112 0

View Details

Kostas @christidis

2 days ago

@athyuttamre @athyuttamre bumping this for when you get a chance. Is my understanding correct that Voice on iOS Remote works on a single persistent chat? As opposed to each session creating a new task/chat.

0 0 0 7 0

View Details

Kostas @christidis

5 days ago

@athyuttamre Starting Voice from iOS Remote with no task selected resumes an old voice task instead of creating a new one. Bug?

1 0 0 21 0

View Details

nic carter @nic_carter

2 days ago

With some very minor exceptions, the data center buildout is happening in red states (historically some purple but on a go forward basis, red). The data center question is entirely a red on red issue. It’s impossible to build in blues and power is expensive, hence no one is even planning to build there 🚨this is why the left is poisoning the well on datacenters 🚨 They have nothing at stake. Nothing to gain, nothing to lose. They will run the Greta / fracking playbook on DCs. The right must avoid being useful idiots for the left and China at all costs. By all means, debate the merit of the DC buildout. But don’t fall into the slopulist trap of being against them “because everyone is” or because the left manufactures consent via the press and NGOs. The left is not participating in the DC buildout; they should not have a seat at the table. Assume that they will try to play spoiler to ruin what is a massive reindustrialization opportunity in Americas heartland, creating skilled jobs that have been missing for so long. It’s entirely rational for the left to ruin this for red America. Don’t let them.

194 463 3K 225K 604

View Details

Kostas @christidis

3 days ago

My experience with Pangram suggests it's best thought of as a reverse Bloom filter; a "yes" is definitely yes, but a "no" is maybe. You can argue that folks only post the 100% AI screenshots (they're using the Bloom filter the right way!), but my knee-jerk reaction is still, "do you realize how trivially this can be gamed"

Byrne Hobart @ByrneHobart

3 days ago

Pangram is good, but has some well-known failure modes, like: - A time you heard it didn't work - A school essay you wrote years ago that it flagged as AI (you don't have a link or the essay text handy) - A different AI detector got it wrong, and they're all the same, right?

38 49 1K 48K 85

0 0 1 166 0

View Details

Kostas @christidis

3 days ago

@ChrisGPT I've never got anyone from OpenAI to even acknowledge a bug or a feature request. One day, the streak will break. From last week alone: x.com/christidis/sta… x.com/christidis/sta… x.com/christidis/sta… x.com/christidis/sta…

Kostas @christidis

a week ago

@thsottiaux The Scheduled page should show the model/reasoning effort used for each task.

0 0 0 151 0

0 0 0 92 0

View Details

Kostas @christidis

3 days ago

@GrantSlatton @__drewface the theory-of-mind instruction for the comments is so clever

1 0 3 34 1

View Details

Kostas @christidis

3 days ago

@siegerts can you help?

0 0 0 27 0

View Details

Kostas @christidis

4 days ago

@OpenAIDevs I think there's a bug on the iOS app. The list of remote projects is not being kept up-to-date with the desktop.

1 0 0 67 0

View Details

Kostas @christidis

3 days ago

How you do anything is how you do everything. The fact he couldn't be bothered to right-click > Learn Spelling on "AlphaFold" tells me everything I need to know about this team's attention to detail. Pass.

Jeff Dean @JeffDean

4 days ago

We created a pitch deck to tell a handful of VC firms about us and what we were up to (a fun experience!). Here’s a few slides about our background and some of the things we’ve worked on from the pitch deck (it was fun putting together the list of people in our teams who have

241 413 7K 3.4M 3K

0 0 1 211 0

View Details

Kostas @christidis

3 days ago

@nikitabier Best to ever do it. Thanks for your great work.

0 0 0 8 0

View Details

Kostas @christidis

3 days ago

@RobinhoodApp been a day one, and a $HOOD investor as well

0 0 0 75 0

View Details

Ben (no treats) @andersonbcdefg

4 days ago

google should randomly execute 1 deepmind employee per day until they can make a model better than kimi k3. this is starting to get embarrassing

99 83 4K 645K 236

View Details

Kostas @christidis

4 days ago

@C_Chichester thank you!

0 0 0 10 0

View Details

Kostas @christidis

4 days ago

Unsolicited endorsement: been using @maticrobots for a while now, and it's worth every penny. It is the Tesla of robo-vacuums; a delightful product which gets even better with every software update.

1 1 9 815 0

View Details

Kostas @christidis

4 days ago

I realized pretty quickly that nobody in that industry knows what they're talking about. Babies survive anyway because (a) most of the advice is innocuous and (b) they're resilient. Also, @xwanyex's post about the bottom quintile – a surprising number of warnings are just dead obvious if your IQ is above 85.

Mike Bird @Birdyword

4 days ago

Really funny to encounter the newborn baby advice bullshit industry firsthand. You're immediately inundated with contradictory pieces of advice from professionals, and when you go to look it up you often find neither side actually has any evidence to back them up at all.