Indeed paused its "Apply for Me" auto-apply feature after employers and recruiters pushed back on bot-generated application floods.
Volume-first auto-apply made the noise problem worse, not better.
Quality matching isn't a nice-to-have. It's the only thing that works rn.
Every country except one lets a model reject you silently.
The EU fixes that on Dec 2, 2027 recruitment AI reclassified as high-risk, with audits and human oversight attached.
That’s 15 months of black box, and only for Europeans.
#Aijob
My job scanning agent found a 75% match this morning.
Then it told me the application was behind a membership paywall.
An agent that only reports good news isn’t an agent. It’s a slot machine.
Three others running: content, announcements, DevOps briefing
254 applicants per job ad. Not because 254 people were qualified. Because applying is now nearly free. The cost of applying collapsed, so the value of applying collapsed. Stop adding to the pile. Start finding the jobs actually worth applying to
. @Greenhouse processes 22 million job applications a month.
Yet 75% fewer of them make it to the hire stage than 18 months ago.
CEO Daniel Chait coined a term for it: the AI doom loop. He came on @chain_ofthought to provide a roadmap for technical job seekers, and break down why we're seeing more applications than ever before, yet hiring has gotten harder - nobody is happy in today's job market.
Plus learn what's replacing leetcode and the case for killing the resume entirely 👇
Chapters:
00:00 Cold open: 239% more applications, 75% fewer hires
03:05 The AI doom loop, defined
04:01 How we got here: remote work, ZIRP, and ChatGPT
07:51 Are software engineering jobs really in trouble?
12:46 The trust crisis: 91% of recruiters spot deception
15:52 North Korean spies, deepfakes, and laptop farms
19:34 Can AI fix the problem it created?
20:52 AI screener interviews and the uncanny valley
26:33 Greenhouse Dream Jobs: one signal, 5x conversion
28:31 Why auto-apply doesn't work (and what does)
30:18 Communities, building in public, and the early-mover advantage
37:08 Gen Z lost trust, and the bias problem
39:04 Kill the resume: rethinking hiring from scratch
43:34 How Greenhouse changed its own interview process
48:47 Coding interviews in the agent era: leetcode is dead
51:33 Predictions: more proof, more conversations, less noise
54:34 Where job seekers and hiring teams should start
A browser that renders nothing a human would ever look at.
Cloudflare’s Kitesurf: 3-7x less CPU and memory than Chromium, ~1.7x slower. Agents don’t care about slow. They care about cost per session.
The browser was always shaped like a person. That constraint just became legacy
Introducing Kitesurf: a browser built for agents, running entirely on Cloudflare Workers.
Chromium is too heavy to hand every agent one. Kitesurf is written in Rust, uses 3-7x less CPU and memory, and spins up per request.
Free in beta: https://t.co/Vm8kVCgSkG
The interesting part isn’t that Claude watermarks its output. It’s that a rewrite removes it.
So the mark filters out the lazy, not the deliberate. Provenance you can launder with effort.
We just built the world’s most polite honesty system. 😎
Amazon now lets AI agents run continuously for up to 14 days (AgentCore on EC2).
Chat was phase one of AI. Long-running work is phase two.
We are moving from asking AI questions to employing it.
Source: https://t.co/IkKZN95XlT
OpenAI paused development of its upcoming Astra model after internal evaluations suggested it may have crossed its own "critical" cybersecurity threshold: autonomously finding and exploiting serious vulnerabilities without human help.
@Reuters
Some days I genuinely don’t feel like building.
Not because I’ve lost belief in what I’m working on. Sometimes the brain just says, “not today.”
Then you sleep, wake up, take a shower, and somehow the problems look smaller.
Some of the biggest companies automated first and asked questions later. Now they are quietly rehiring people they let go, because the automation did not deliver. The lesson is not that AI failed. It is that speed without understanding is expensive.
The most underrated AI skill right now is not even prompt engineering anymore
It is knowing what not to automate.
Automate the boring step. Keep the human in the loop where judgement, trust, or accountability matters. That line is where real productivity gains live
Vibe coding is more addictive than video games.
You build something → it works → you get an idea → you build that → suddenly it’s 3AM and you’re like:
“Wait… I can actually build anything now.” 😭
The dopamine loop is REAL bro