Ajmal Khan

🦺 🛟 Agent Safety > Model Safety LLMs didn't break your app. Agents can. As we wire LLMs to tools, browsers, payments & code the risk shifts from bad text to bad actions. Top failure modes (seen in the wild & in papers): 🚩 Prompt/Content Injection -> Tool Misuse. Malicious web pages or data instruct agents to steal secrets or make unintended API calls. This is now a primary risk for browsing and MCP - style tool agents. 🚩 Over permissioned tools & long-lived tokens. Agents get "god mode" scopes. A single injection becomes account takeover or irreversible ops. Security notes for MCP emphasise least privilege and short-lived auth. 🚩 Unsafe web autonomy. Benchmarks show web agents will attempt harmful tasks unless constrained (posting misinformation, illicit sales, etc.). You must measure misuse not just assume guardrails. 🚩 Supply-chain & retrieval poisoning. Agents trust plugins, third-party tools & indexed data that an attacker can taint. OWASP’s newest GenAI Top 10 calls this out explicitly. 🚩 Process safety gaps. NIST's GenAI Profile (AI 600-1) warns that organizations ship agents without role clarity, human-in-the-loop or incident playbooks. ✅ Model safety reduces bad text. Agent safety prevents bad state. Treat the agent like a junior SRE - with narrow permissions, audited actions and clear escalation paths. If you're building agents, what's the one control you won't ship without? Also drop a comment if you'd like me to share a follow-up post on how to tackle each of these risks. #AgenticAI #AISafety #AIAgents #GenAI #CyberSecurity #MCP #OWASP #NIST #AITrust #ResponsibleAI #AIForBusiness

5

Does this also mean we’re entering a phase where VCs and public markets investors approach companies with a predefined mindset around revenue per employee or optimized team as a key metric, Deedy Das?

3

Trend: AI Salaries Splitting the Market 📉 The software engineering market is splitting into two. 💸 Top AI/ ML engineers → $1M - $2.5M total comp (OpenAI, Anthropic, Meta) ⚖️ Median SWE offers → flat or even down YoY 📈 Layoffs still happening in parallel It's looking more like pro sports: 🏆 The top 1% get record-breaking contracts ⚔️ The rest are fighting for fewer spots as AI tools automate parts of the stack Here’s the blind spot: 👉 Most AI startups are fighting over the same US-based talent pools… 👉 While regions like Eastern Europe (with huge engineering depth) remain undervalued ⚡Startups will either need to: 1️⃣Overpay for elite engineers, or 2️⃣Bet on global/ undervalued talent pools ⚡ The real winners won’t just overpay for superstars They’ll build async, remote-first teams, tapping into global talent pools that others are still ignoring 🕵‍♀️ Sharing a recent snapshot from Levels.fyi - the top 10 software engineering offers in the US right now #AI #Startups #TechHiring #RemoteWork #Compensation #SoftwareEngineering

29

1 Comment

The $450k vs. $84 Debate: What are we actually paying for in 2026? 🧠 The question is no longer "Will AI code?" but "What are we actually paying software engineers for?" If you compare a Senior SWE (L5) at a FAANG company to frontier models like GPT-5.2 or Claude 4.5, the cost-per-token delta has moved from "massive" to "astronomical." 📊 The Annual Breakdown (Human vs. Machine): 👤 Human SWE (L5): **$450,000** | ~20M tokens | Cost per 1M: **$22,500** 🤖 AI (GPT-5.2): **$84.00** | ~20M tokens | Cost per 1M: **$4.20** 🤖 AI (Claude 4.5 Sonnet): **$180.00** | ~20M tokens | Cost per 1M: **$9.00** *Sources: Levels.fyi (Meta E5 total comp approx $450k), OpenAI API Pricing ($1.75 in / $14.00 out), Anthropic API Pricing ($3.00 in / $15.00 out).* To understand the 5,000x premium, we have to look at "Cognitive Throughput." A human developer doesn't just read—they process context, history, and system architecture. A human reads at ~300 words/min (roughly **400 tokens/min**). Add in thinking, debugging, and meetings, and a "Thought-Hour" equals about 10,000 tokens. An AI processes those same 10,000 tokens in about **150 seconds**. 💡 So why pay $450k? Because tokens aren't judgment: 1️⃣ Context Density: Humans manage "Implicit Context"—the stuff NOT in the README/Claude.md (vast experince of SDLC scenarios, legacy debt, 5-year visions etc). 2️⃣ Accountability: An LLM doesn't care if the production database drops at 3:00 AM. A human does. (This is debatable, but given the current progress still needs human judgement) 3️⃣ Problem Definition: AI is world-class at solving problems, but humans are the only ones who know which problems are worth solving. The most valuable engineers today aren't competing with the dollar-per-token API cost. They're using $84 of compute to multiply their $450,000 of judgment. The arbitrage isn't in replacing the human—it’s the 100x leverage of your "Thought-Hour." 🚀 #SoftwareEngineering #AI #FutureOfWork #TechTrends #OpenAI #Claude

5

Imagine you just joined Google as an SWE. It’s your first week, and you’re overthinking everything. You’re surrounded by world-class engineers, new systems, and an overwhelming urge not to mess up. The pressure to prove yourself is real, and so is the anxiety…so what should you do? Over the years, I’ve moved across orgs, teams, and sometimes, entire companies. Here’s exactly what I did (and still do) to ramp up faster and enjoy the process without letting anxiety win → Don’t let yourself stay stuck. If you’re blocked for more than 30 minutes, ask for help. Share what you’ve tried, show initiative, and intent. → Read way more code than you’re assigned. Dive into related modules, historic PRs, and old design docs. The context you gain helps you spot patterns and avoid mistakes. → Ask about the story behind the system. Find out why things are built a certain way, what’s failed before, and what’s on the roadmap. Understanding the “why” beats memorizing the “what.” → Volunteer for unglamorous work. Take ownership of tests, docs, or nagging integration bugs. This is where you quietly learn the most and build your internal network. → Offer help before you’re an expert. Chime in when you see questions, review PRs, or share a script. Even small contributions get you noticed for the right reasons. → Connect with skip-levels and cross-team leads. Set up short intros to learn about the broader vision and team priorities. You’ll spot opportunities to add value beyond your immediate tasks. → Document what you learn as you go. Share notes, update READMEs, or post quick how-tos in team channels. Your future self and your teammates will thank you. → Shadow someone during incidents or launches. Sit in on debugging sessions or war rooms, even if you’re just observing. You’ll see how senior folks think, react, and communicate under pressure. → Keep a running list of “how did that break?” moments. Every time you see a system fail or a hotfix go live, dig in. Reverse-engineering past problems is the fastest way to understand the architecture. → Show up with curiosity and intent, every day. Don’t just tick off JIRA tickets. Look for the bigger picture. People notice when you genuinely want to learn and contribute, not just “do your job.” Some orgs assign mentors, some don’t. Either way, these habits put you in the driver’s seat. Ramping up is less about brilliance, more about consistent intent. And the ones who learn this early end up leading teams later.

709

51 Comments

Entry-level SWE: $80K. Senior (5-7 years): $360K. Staff at FAANG: $500K-$900K+. Whether you started at $70K or $90K won't matter in 5 years. The real cost of holding out: 6 months of unemployment = $31K in lost wages. Holding out for an extra $10K/year while unemployed for 6 months? You're down $21K net. What one year of experience unlocks: Job hoppers see 15% average raises when switching. Some see 20-30%. Start at $80K. Work one year. Switch jobs with a 15% raise. You're at $92K. That's more than the $90K offer you turned down. What to focus on: 1. Get any legitimate SWE role 2. Learn as much as possible in year one 3. After 12-18 months, switch for a 15-20% raise The exception: Don't accept insultingly low offers. If market is $90K and someone offers $50K, walk away. But $75K vs $85K? Take whichever starts sooner. Are you holding out for the perfect first offer, or are you getting started?

53

2 Comments