Thousands of Zero-Days, 30K Oracle Cuts, and Meta's Muse Spark

This week, Anthropic announced a model that autonomously discovered thousands of zero-day vulnerabilities in every major operating system and browser — then decided not to release it publicly. Meanwhile, Oracle quietly cut up to 30,000 jobs to fund AI data centers, Meta debuted its first model from its new Superintelligence Labs, and Jeff Bezos poached an xAI co-founder for his $6.2B AI startup. Edition #50 is a milestone week, and the signals are loud: security hiring is about to surge, the AI talent war is escalating to the C-suite, and the line between "AI company" and "every company" just disappeared.

The Drop

Anthropic launches Project Glasswing after Claude Mythos finds thousands of zero-days

What happened: On April 7, Anthropic revealed Claude Mythos Preview, a model so capable at cybersecurity tasks that it autonomously identified thousands of previously unknown vulnerabilities across every major OS and browser. Rather than releasing it publicly, Anthropic launched Project Glasswing — a restricted initiative providing Mythos access to roughly 40 organizations including AWS, Apple, Microsoft, Google, and CrowdStrike to help secure critical infrastructure.

Why it matters for hiring: This is the clearest signal yet that AI-driven security is moving from experimental to essential. Every organization in the Glasswing partnership will need teams who can work alongside autonomous vulnerability-discovery tools. The demand for security engineers who understand AI-assisted threat detection is about to spike hard. Companies outside the partnership will scramble to build defensive capabilities against the same class of AI-powered attacks.

Roles likely to spike:

• AI security engineers and red team specialists

• Vulnerability researchers with ML/AI fluency

• Security operations engineers (SOC automation)

• Application security architects

Oracle cuts up to 30,000 jobs to fund $300B AI data center buildout

What happened: Oracle has laid off more than 10,000 employees globally, with estimates indicating total cuts could reach 30,000 — the company's largest restructuring ever. The layoffs span cloud, healthcare, sales, and NetSuite operations. Oracle cited AI-driven code generation enabling operations with smaller teams, and the savings are being redirected toward a $300 billion cloud infrastructure deal with OpenAI.

Why it matters for hiring: This is the clearest case of "cut to fund AI" at enterprise scale. Oracle's move creates a massive pool of experienced cloud, ERP, and enterprise software talent hitting the market simultaneously. For hiring teams at mid-market companies, this is a rare window to pick up senior enterprise talent that's normally locked behind Oracle's comp packages. But it also signals that traditional enterprise roles are compressing — teams that can't demonstrate AI-augmented productivity are vulnerable everywhere.

Roles likely to be available:

• Cloud infrastructure and platform engineers

• Enterprise SaaS and ERP specialists

• Healthcare IT and data integration engineers

• Sales engineers and solution architects

Meta debuts Muse Spark, first model from Superintelligence Labs

What happened: On April 8, Meta unveiled Muse Spark — the first AI model from Meta Superintelligence Labs, the unit created after Mark Zuckerberg's $14.3 billion deal to bring in Scale AI's Alexandr Wang. The model is natively multimodal with tool-use, visual chain of thought, and multi-agent orchestration capabilities. Benchmark results show it's competitive with GPT-5.4 and Claude Sonnet 4.6 across most tasks.

Why it matters for hiring: Meta's hiring spree to staff Superintelligence Labs pulled executives from OpenAI, Anthropic, and Google — and they're not done. The launch of Muse Spark validates Meta's AI pivot and signals continued aggressive talent acquisition. For recruiters, this means top-tier AI research talent is even harder to retain, and comp packages at the frontier labs keep escalating. Meta also created a new applied AI engineering org under Maher Saba, meaning applied ML roles at Meta are expanding rapidly.

Roles likely to spike:

• Multimodal ML researchers

• AI agent and orchestration engineers

• Applied AI/ML engineers

• AI safety and alignment researchers

Bezos' Project Prometheus poaches xAI co-founder in escalating talent war

What happened: Jeff Bezos' $6.2 billion AI startup Project Prometheus hired Kyle Kosic, a co-founder of Elon Musk's xAI and former senior OpenAI executive, to lead infrastructure efforts. Kosic previously helped build xAI's Colossus supercomputer. Prometheus, focused on applying AI to manufacturing, aerospace, and computing, now has 120+ employees poached from Meta, OpenAI, and DeepMind.

Why it matters for hiring: This is talent-war escalation at the highest level. When billionaire-backed startups start pulling co-founders from other frontier labs, the ripple effects cascade through the entire market. Every AI infrastructure lead who leaves creates a backfill need, and the replacement pool at this level is vanishingly small. The signal for hiring teams: if you're recruiting AI infrastructure talent, expect longer cycles, higher comp demands, and candidates fielding multiple offers from deep-pocketed competitors.

Roles likely to spike:

• AI infrastructure and systems engineers

• Large-scale compute/cluster architects

• Robotics and physical AI researchers

• ML platform engineers

Eclipse Ventures closes $1.3B for physical AI — the largest "build from scratch" bet yet

What happened: Eclipse Ventures closed $1.3 billion across two funds targeting physical AI — the convergence of AI with robotics, autonomous systems, and hardware. The firm will use a portion to incubate startups from scratch, not just invest. Portfolio includes Wayve (autonomous vehicles), Cerebras (AI chips), and Bedrock Robotics (self-driving construction vehicles). Total AUM now exceeds $10 billion.

Why it matters for hiring: The physical AI thesis is attracting serious capital, which means robotics, autonomous systems, and AI-hardware hybrid roles are about to see sustained demand. Eclipse's "venture building" model means new startups will appear needing founding engineering teams. For recruiters, this is an early signal to start building pipelines in robotics, embedded systems, and AI-for-manufacturing before the market heats up further.

Roles likely to spike:

• Robotics engineers (perception, planning, control)

• Embedded systems and firmware engineers

• Autonomous vehicle software engineers

• AI hardware/chip design engineers

AI Tool of the Week

LinkedIn Hiring Pro AI Interviews ("AI-assisted screening interviews at scale")

What it does: LinkedIn's new AI interview feature, now in early testing for Hiring Pro users, lets employers invite top applicants to complete audio or video screening calls with an AI interviewer. The AI generates questions based on the job description, conducts the interview, then delivers transcripts, recordings, summaries, and a 5-point rating to recruiters. Hirers can manually select up to 40 candidates or let LinkedIn auto-identify top applicants from the pool.

Who it's for: Hiring managers and recruiters at small-to-mid-sized businesses who need to screen high volumes of applicants without burning recruiter hours on first-round calls.

Quick pilot idea (this week):

• Pick one high-volume role currently open (ideally 50+ applicants)

• Set up a Hiring Pro AI screening with 5-7 role-specific questions

• Run 20 candidates through the AI screen over 5 days

• Compare AI ratings against your team's manual screening decisions on the same candidates

Metrics to track:

• Time saved per candidate vs. traditional phone screen

• Correlation between AI rating and hiring manager interview pass rate

• Candidate completion rate (how many invited actually finish)

• Candidate experience feedback scores

Hiring / Interview Insight

Your bottleneck is scheduling, not sourcing

New data from GoodTime's 2026 Hiring Insights report shows that recruiters spend 38% of their time scheduling interviews — the single biggest operational tax measured. 42% of candidates drop out when scheduling is slow, and teams using automated scheduling are 1.6x more likely to hit their hiring goals.

Meanwhile, recruiters are handling 93% more applications and 40% more open reqs than in 2021 — with 14% smaller teams. The math doesn't work without automation.

One change to implement:

• Audit your scheduling workflow this week. If any step requires a recruiter to manually coordinate calendars, replace it with automated scheduling (Calendly, GoodTime, ModernLoop). Target: reduce median time-from-screen-to-onsite by 2 business days within 30 days.

Funding Watch

Shield AI | $2B | Series G at $12.7B valuation
Likely hires: autonomous systems engineers, simulation engineers, defense AI researchers, flight software engineers

Project Prometheus (Bezos) | $6.2B | Launch round
Likely hires: AI infrastructure engineers, robotics researchers, manufacturing AI specialists, compute architects

Eclipse Ventures | $1.3B | New fund (physical AI)
Likely hires: robotics engineers, embedded systems developers, autonomous vehicle engineers, AI hardware designers

Halter | $220M | Series E at $2B valuation
Likely hires: IoT/embedded firmware engineers, ML engineers (edge AI), field operations managers, agricultural data scientists

Deepgram | $130M | Series C at $1.3B valuation
Likely hires: speech/NLP engineers, voice AI researchers, developer relations, enterprise sales engineers

Cyberhill Partners | $11M | Growth round
Likely hires: enterprise AI solution architects, delivery engineers, go-to-market leads

Quick Bytes

• Q1 2026 broke every VC record — $300B invested globally, with AI accounting for 80% ($242B). Four companies alone (OpenAI, Anthropic, xAI, Waymo) absorbed $188B. The concentration means most AI hiring demand flows through a handful of mega-funded orgs.

• 78,557 tech workers laid off in Q1 with nearly 48% of cuts attributed to AI automation. IBM bucked the trend by tripling entry-level hiring, arguing AI still needs human oversight. If you're losing candidates to "AI will replace this role" fear, IBM's stance is a useful counter-narrative.

• GPT-5.4 launched with native computer-use — OpenAI's latest model can operate computers and carry out complex workflows across applications. Expect demand for "AI ops" roles (people who design, monitor, and govern autonomous AI workflows) to accelerate sharply.

What to do this week

1) Build a security hiring pipeline now
Action: Identify 3-5 security engineers in your network who have AI/ML crossover skills. Reach out with a market-update conversation, not a job pitch. The Glasswing announcement means this talent segment is about to get very expensive very fast.
Metric: Number of warm security+AI conversations started this week.

2) Source from Oracle's talent pool
Action: Set up LinkedIn alerts for Oracle alumni in cloud, platform engineering, and enterprise SaaS roles. Many of these professionals are senior, well-credentialed, and newly available. Move fast — the best ones will be off-market within 4-6 weeks.
Metric: Number of Oracle alumni contacted and response rate.

3) Audit your interview scheduling stack
Action: Map every step between "candidate passes screen" and "onsite interview scheduled." Identify manual handoffs and replace at least one with automation this week.
Why: 42% candidate drop-off from slow scheduling is a fixable leak in your funnel. Every day you shave off time-to-onsite improves your close rate.

That's edition #50, and what a week to hit the milestone. The biggest takeaway: AI capability is accelerating faster than hiring teams can adapt. Models are finding zero-days, generating code that replaces teams, and operating computers autonomously. The companies that win talent in this market are the ones that move fastest — in scheduling, in outreach, and in recognizing which roles are spiking before the rest of the market catches up. Stay informed, and see you next Monday!