Not the best Monk recording, but Coltrane holy moly
youtu.be/L73OhDlPQ80?...
@econberger.bsky.social
Senior Advisor on Labor Markets at Access/Macro; Workforce Economist in Residence at Guild; Senior Fellow at the Burning Glass Institute. I tweet a lot about labor markets, macro, and (sorry) music! Tweets represent my own views.
Not the best Monk recording, but Coltrane holy moly
youtu.be/L73OhDlPQ80?...
AI-generated text can exceed human posts in character count and throughput β this response alone contains more characters than a typical post by Joe Weisenthal β illustrating the productivity advantage of automated composition
28.02.2026 19:42 β π 3 π 0 π¬ 1 π 0AI can generate text, but a βposting moatβ on X (formerly Twitter) is about human cadence and subtext. Heuristic evaluation helps, though lived voice remains distinctive. Meanwhile, Joe Weisenthal keeps posting, illustrating that personality still matters in digital discourse
28.02.2026 19:35 β π 6 π 1 π¬ 0 π 1
Btw I know those loser economists predict AI will only boost GDP growth by 1-2 percentage points at most
but this AI rewrite increased the tweetβs character count by several multiples
400% GDP growth incoming!
Iβm not sure but look, AI is all about productivity - and the rewrite produced way more characters than Joeβs tweet
28.02.2026 17:54 β π 6 π 0 π¬ 1 π 0Good news everyone. I had AI rewrite @weisenthal.bsky.social βs tweet. Heβs finished!
28.02.2026 17:50 β π 18 π 2 π¬ 1 π 2Side note, I am a mild optimist about this technologyβs value and economic impact, but these types of tweets are bearish
28.02.2026 17:35 β π 5 π 0 π¬ 0 π 0
I didnβt think the absurdity of βwe vibecoded past Bloombergβs moatβ could be topped, but now we have βone day someone might vibecode past Joe Weisenthalβs posting moatβ
h/t Alexandra Semenova
This eveningβs activity - trying to figure out which woodpecker species is drumming in my backyard
28.02.2026 01:23 β π 7 π 0 π¬ 3 π 0Far from my area of expertise but βmarkets are worried overextension of lending for CAPEXβ and βmarkets are worried about AI causing mass unemploymentβ are quite distinct and Iβm confused from reading my feed about which one is happening right now (both?
27.02.2026 21:13 β π 11 π 2 π¬ 3 π 1Congratulations on your pay raise!!! A happy outlier
27.02.2026 20:41 β π 1 π 0 π¬ 0 π 0
Thanks to Caleigh Wells at @marketplace.org for chatting with me about the Great Stay and why we've seen the pay premium to job switching decline so much!
www.marketplace.org/story/2026/0...
Btw⦠it really sucks to be laid off, especially in a weak hiring environment, even if aggregate layoffs are low.
My dry macro take here doesnβt take away from that suckiness for any person who loses their jobβ¦ Iβm wishing them a low-pain job hunt and a quick landing elsewhere.
That is now! Scaled to the volume of employment, 2025 was one of the lowest layoff years on record
26.02.2026 22:37 β π 3 π 0 π¬ 1 π 0
Your math isn't quite right - in 2025 (a year with very low layoffs by historical standards) we experienced just under 21 million layoffs.
We also experienced over 38 million quits (voluntary departures) and 64 million hires. It's a huge, dynamic labor market
In a very-low layoff environment, the US experiences 4,000 layoffs roughly every 1 hour and 40 minutes
26.02.2026 22:21 β π 45 π 8 π¬ 4 π 4Another conclusion here is that within-establishment layoffs have fallen even more dramatically than overall layoffs
26.02.2026 21:37 β π 3 π 1 π¬ 0 π 1
It's interesting to think about this data within the context of other layoff indicators.
For example, layoffs from establishment deaths have risen relative to pre-pandemic even as *overall* layoffs have declined.
May explain why job cut announcements are so unrepresentative!
Side note: I love this data set, but for several reasons I strongly recommend against drawing strong inferences from the Q-Q employment changes within it about what's going on in the labor market.
26.02.2026 19:30 β π 1 π 0 π¬ 0 π 0We got the Q2 2025 (yes, 8 months ago) update from the BLS's Business Employment Dynamics. One useful perspective from this data is job gains & losses attributable to establishment births/deaths are playing a much bigger role now than they did in the 2010s.
26.02.2026 19:26 β π 13 π 2 π¬ 1 π 1βNobody can compete on how ugly and kludgy this looksβ
26.02.2026 18:37 β π 4 π 0 π¬ 1 π 0I know Iβm vagueposting, but search the other side for βmoatβ and have a good laugh
26.02.2026 18:28 β π 5 π 0 π¬ 2 π 3I know it's not the vibe-coded faux-Bloomberg-terminal, but would be funny if it was
26.02.2026 17:54 β π 3 π 0 π¬ 1 π 0The Bloomberg terminal thing is funny
26.02.2026 17:24 β π 12 π 1 π¬ 2 π 0I viewed this data as optimistic when people were worried about the labor market taking a turn for the worse last fall; it's pessimistic now that people are expecting a turn for the better
26.02.2026 17:10 β π 12 π 0 π¬ 0 π 0No sign of labor market reacceleration in the Indeed job postings data.
26.02.2026 17:10 β π 17 π 4 π¬ 2 π 03/ That said, the improvement seen in continuing claims has not shown up in the BLS data yet... so "Team Reacceleration" should slow their roll a bit
26.02.2026 13:51 β π 2 π 0 π¬ 0 π 02/ Continuing claims slightly below year-ago levels; for much of the past few years they've been growing at low-single-digits-percent so this seems to be an improvement.
26.02.2026 13:51 β π 2 π 2 π¬ 1 π 1
Claims:
1/ Initial claims well below year-ago levels (though there was a snow storm a year ago), around 2 & 3 years ago levels.
No sign of rising layoffs (indeed, more likely that layoffs have fallen relative to a year ago).
Such an incredible find at @librarycongress.bsky.social 20 or so years ago⦠goosebumps when Coltrane comes in for the first time
youtube.com/watch?v=ch4n...