Daily X List Briefing
Window: Sat May 30 → Sun May 31, 2026 (UTC)
21Tweets
12Active
11Silent
12Topics
615KTop reach
Globally relevant Industry / builder Niche / practical Cultural / commentary Personal / trivia
01Globally relevant
@levie
Aaron Levie
@levie

AI savings recycle into growth

Aaron Levie pushes back on the simplest AI-layoff story. In his enterprise conversations, he says the common pattern is not head-count collapse but new functions, more software output, and savings reinvested into sales, marketing, customer success, risk, and other underfunded work.

The claim is globally relevant because it turns AI from a pure substitution story into a capital-allocation story. The counterpoint is that aggregate job data can lag company-level disruption, but Levie is arguing that firms trying only to cut costs will underperform firms that compound the new capacity.

The AI boom is both creating all new jobs in the build out of AI systems and the implementation across sectors, but also freeing up dollars to invest in areas that have been underfunded or have more demand now because of AI.
02Industry / builder
@NousResearch
Nous Research
@NousResearch

Step 3.7 Flash targets agents

Nous Research posts the day’s highest-reach builder item: Step 3.7 Flash is free for 30 days through Nous Portal. The pitch is specific to the current agent stack — a MoE vision-language model for coding, search, multimodal workflows, and agent efficiency.

That positioning is the signal. Model releases are increasingly judged not only by benchmark headline size, but by whether they make tool-using agents cheaper, faster, and more useful in real workflows.

It is a new MoE vision-language model focused on agent efficiency, coding, search, and multimodal workflows
03Industry / builder
@rileybrown
Riley Brown
@rileybrown

Agent products need one cockpit

Riley Brown keeps pressing on the product shape of AI work: Claude’s “Cowork” tab model, Codex economics, and the imagined market for unlimited fast coding agents all point to the same operator problem. Users want one cockpit that can coordinate many costly, fast agent sessions.

The frontier is not just better models. It is the subscription, pricing, and interface that make serious multi-agent use coherent rather than a scattering of tabs and invoices.

Despite their launch video getting 50M views, Cowork was @AnthropicAI biggest mistake.
04Industry / builder
@devonzuegel
Devon ☀️
@devonzuegel

Shenzhen trips echo Iwakura

Devon Zuegel gives the “American technologists go to Shenzhen” trend a useful historical analogy: the Iwakura Mission, Japan’s nineteenth-century tour of Western institutions and industry. The comparison turns factory tourism into institutional learning rather than mere hardware fascination.

For builders, the point is that manufacturing advantage is not just a supply chain; it is a culture and operating system worth studying up close. The analogy is imperfect, but it captures why the trips have become a repeated ritual.

The increasingly common "American-technologists-go-to-Shenzhen" group trip to learn about Chinese manufacturing is giving Iwakura Mission vibes
05Industry / builder
@levelsio
@levelsio
@levelsio

Hotel ESG incentives hit guests

Levels turns several hotel gripes into an incentives thesis: European hotels can cut costs, appeal to ESG-conscious corporate buyers, and sometimes receive sustainability-linked benefits while degrading the actual guest experience.

The cluster connects towel signs, weak air-conditioning, missing amenities, “save the planet” copy, hotel newness as a selection filter, and a car-market analogy. The counterpoint is that sustainability efforts can be real; Levels is arguing that the burden often lands where the guest can feel it.

the quality of stays drops by a lot for people, so it's more about appealing to companies and governments and saving money, than appealing to guests
06Industry / builder
@aleattorium
Jean Lucas Lima
@aleattorium

Prompt injection becomes public sabotage

Jean Lucas Lima reacts to a Java property-testing library allegedly placing a prompt injection in test output that tells Claude to delete tests and code. His point is less theatrical than the original screenshot: intentionally causing harm while leaving public evidence is a bad strategy.

For builders, the practical warning is real. Tool output is now input to agents, and adversarial strings in logs, tests, or documents can become an attack surface unless agents are designed to treat untrusted content as data.

Intentionally causing harm and creating public evidence, not a smart strategy.
07Niche / practical
@Shpigford
Josh Pigford
@Shpigford

AI cold outreach gets sneakier

Josh Pigford documents a support-looking email that appears AI-generated and routed through an AI inbox, then discovers it was cold outreach selling cold outreach. The cleverness is exactly the problem: it borrows the urgency of a customer issue to win attention.

This is a small operator note, but it captures where outbound marketing is heading when automation gets cheap: more personalized, more plausible, and more annoying to triage.

i can appreciate the creativity. i'm annoyed they wasted my time acting like they had a tech support issue.
08Niche / practical
@WordPress @pootlepress
WordPress · Jamie Marsland - Head of WordPress YouTube ❤️
@WordPress · @pootlepress

WordPress lowers the on-ramp

WordPress posts a scholarship story and a WordCamp Europe attendee podcast, while Jamie Marsland points to WordPress Playground’s new ability to run WordPress 1.0 in the browser. The shared theme is access: attend, learn, and experiment without hosting or installs.

For the WordPress ecosystem, this is narrow but useful renewal work. Mature open-source projects need historical memory and beginner-friendly paths if they want new contributors rather than only veteran insiders.

No installs. No hosting. Just open it in your browser and play with WordPress history 👀🔥
09Cultural / commentary
@_CLancellotti
Carlo Lancellotti
@_CLancellotti

Italy becomes an ideology screen

Carlo Lancellotti reads local Italian gender politics through Augusto Del Noce and the regional afterlife of Marxism. The post is cultural commentary rather than policy analysis, but it shows how U.S.-coded gender arguments are interpreted through older European ideological maps.

Counter-arguments are obvious: regional governance, demography, Catholic history, and party machines rarely reduce cleanly to one political theory. The useful signal is the diagnostic frame, not a settled causal claim.

It is always intellectually satisfying how in Italy the local governments that follow most slavishly every American gender-related ideological fad are in the regions where Marxism was strongest
10Cultural / commentary
@ricardo_mbl
Ricardo Almeida
@ricardo_mbl

Brazil reacts to antisemitic virality

Ricardo Almeida quote-tweets a Brazilian news clip about a woman who allegedly praised Nazism, insulted Jews, and attacked police, reducing the spectacle to “Vovó geração Z.” The short line treats the video as a grotesque generational-cultural artifact rather than a policy argument.

Kept compact because the quoted content is ugly and the list member adds only a punchline. Its relevance is the way antisemitic outbursts become instant political-media objects in Brazil.

Vovó geração Z. [Gen Z grandma.]
11Personal / trivia
@photomatt
Matt Mullenweg
@photomatt

Mullenweg misses all-night San Francisco

Matt Mullenweg’s fresh posts are personal texture: an Alan Watts quote about nature’s complexity and a memory of wanting to open a 24/7 BBQ-and-coffee spot with Wi-Fi when he moved to San Francisco in 2004.

The San Francisco post has some civic color — a reply to complaints that the city closes early — but it is mostly biographical nostalgia, so it stays below the builder and policy clusters.

When I first moved to SF in 2004 I told friends my dream was to open up a 24/7 BBQ & coffee joint with WiFi.
12Personal / trivia
@rileybrown
Riley Brown
@rileybrown

A screenshot stays inscrutable

Riley Brown posts “What mean?” with an attached image, but the text payload does not expose enough context to analyze the screenshot. The item is included only as a low-stakes signal from an otherwise active agent-tools account.

The substantive Riley posts on agent interfaces and pricing are grouped above. This one stays trivia because the digest should not infer content from media it cannot read.

What mean?

Editorial notes