The AI & Tech Society by Danar

Job postings for Forward Deployed Engineers (FDEs) have surged over the past 18 months, making the role one of the fastest-growing in the tech industry. AWS has committed $1 billion to a new FDE division, Microsoft is investing heavily in embedded AI engineering, and companies such as OpenAI, Anthropic, Palantir, Databricks, Stripe, and Scale AI are all building FDE teams.

This is not just a buzzword. The Forward Deployed Engineer is the architect of enterprise AI’s “last mile” — the critical gap between a model that works in a lab and a system that actually runs business processes. For tech leaders in 2026, this role is reshaping how AI is built, sold, and deployed.

Hosted on Acast. See acast.com/privacy for more information.

1. Tech is the exception — 5.8% vs 3.8% overall; displacement invisible in macro stats
2. Regulation mobilizing — Newsom executive order; EU pressure; state legislation likely 2026-27
3. "More jobs than it destroys" is partly evasive — new roles need different skills; reskilling timeline lags; aggregate doesn't help individuals

1. Be honest about what's changing (no "efficiency" euphemisms)
2. Redirect savings into upskilling, not just GPUs
3. Protect the entry-level rung (new apprenticeship paths)
4. Promote harness skill, not just prompt skill
5. Stop AI-washing organizational decisions
6. Set explicit headcount-vs-AI tradeoffs
7. Treat severance/outplacement as engineering quality

1. Build harness skill, not prompt skill
2. Get certified (e.g., Claude Certified Architect)
3. Track your skill exposure honestly
4. Build a portable, public portfolio
5. Maintain 6-12 months financial runway

1. AI became #1 layoff reason in May 2026 (40%); 7%→40% in five months
2. AI washing is real (6 in 10 companies admit it)
3. The precise truth is capital reallocation
4. CEO statements remarkably consistent (Oracle cut while profitable)
5. Displacement is structural, not uniform (middle hollows out)
6. Tech is the exception (5.8% vs 3.8%)
7. The response defines the next decade

Hosted on Acast. See acast.com/privacy for more information.

1. Adopt a model gateway — centralize routing, failover, governance
2. Build deprecation discipline — retire models deliberately
3. Instrument agents deeply — especially with frameworks
4. Audit prompt caching — fix layout (stable first, dynamic later)
5. Implement budgets & backpressure — cap loops, build queues

1. Multi-model is the norm (70%+ use 3+ models); use a gateway
2. LLM tech debt compounds; retire old models deliberately
3. Framework adoption doubled; observability burden doubled too
4. 69% of tokens are system prompts; only 28% use caching
5. Context windows exploded but quality beats volume
6. Rate limits are the #1 failure mode
7. Agents are still mostly monoliths; distributed shift is coming

Hosted on Acast. See acast.com/privacy for more information.

• February 2026: SpaceX announces xAI acquisition
• Finalized May 6, 2026
• xAI valued at ~$250 billion
• Created vertically integrated "innovation engine"
• Brings Grok, Colossus supercluster, X platform under SpaceX

Hosted on Acast. See acast.com/privacy for more information.

UPDATED ** CLAUDE FABLE JUST GOT SUSPENDED 2026-06-12 BY ANTHROPIC AND THE US GOVERNMENT.

• Fable 5 uses fewer tool calls than Opus-tier models
• 25-30% faster on Anthropic's spreadsheet suite
• Fewer turns partially offset the 2x per-token price
• Measure cost per outcome, not cost per token

Novel design: Routes risky prompts to less capable model rather than refusing

Classifier domains:

1. Cybersecurity
2. Biology and chemistry
3. Model distillation

Fallback model: Claude Opus 4.8 Trigger rate: <5% (Anthropic) / 8-9% (Artificial Analysis) Security testing: 1,000+ hours bug bounty, no universal jailbreak found

Hosted on Acast. See acast.com/privacy for more information.

Key insight: 10.6-point gap on SWE-bench Pro is the largest between Opus 4.8 and GPT-5.5

What it is: Research preview feature letting Claude orchestrate hundreds of parallel subagents

How it works:

1. Claude plans a large task
2. Writes JavaScript orchestration script
3. Spawns tens to hundreds of parallel subagents
4. Runs them simultaneously
5. Verifies results against test suite
6. Returns coordinated final answer

Limits:

• Up to 16 concurrent agents
• Up to 1,000 agents total per run
• "Meaningfully more tokens" than typical sessions
• Available on Max, Team, Enterprise plans

Demonstrated capability: 750,000-line codebase migrated in 11 days with 99.8% test pass rate

Effort LevelUse CaseLowQuick responses, token-efficientMediumBalancedHighDefault for complex workMaxMaximum reasoning depth

Key finding: Opus 4.8 at minimum effort matches Opus 4.7 at maximum effort on SWE-bench Pro

Positive:

• Benchmark gains feel real on agentic coding
• Better on complex, multi-step work
• Proactively flags issues other models miss
• More reliable in long-running sessions

Negative:

• "Wicked Loop of Refactoring" — keeps finding minute issues
• Less legible workings (grep/sed/awk vs edit tool)
• Can get stuck in testing loops
• Misses instructions on simpler tasks
• Worse than 4.7 on some UI generation prompts

Hosted on Acast. See acast.com/privacy for more information.

Panel 1: Vibe Coding

• You → Prompt → Model → Code
• Fast to start
• Feeling over structure
• Good for prototypes
• "You ask the model to solve the problem directly"

Panel 2: What Changed

• Stronger models are not the whole answer
• The new bottleneck is context, rules, and review
• Engineer writes spec → Sets rules → Lets agents work → Reviews output
• "You code less. You steer the system more."

Panel 3: Agentic Engineering

• Agents build. The human orchestrates.
• Bring together: spec, goal, constraints, history, data, rules, tools, tests
• "More scalable. More repeatable. Better results."

Hosted on Acast. See acast.com/privacy for more information.

Stop measuring:

• Lines of code per developer
• Token consumption
• Individual productivity

Start measuring:

• Cycle time (Claude-assisted vs non-assisted PRs)
• Time to first PR for new hires
• PR throughput with quality counterweight (defect rate, rollback frequency)
• Incident resolution time
• Maintenance burden trajectory

Examples from one company:

• Support team: Tool surfacing relevant past tickets and customer history
• Finance team: Expense categorization assistant
• HR team: Onboarding checklist app pulling from live systems

What engineering built:

• Architecture patterns for internal apps
• Plugin marketplace with pre-approved skills/MCP connections
• Managed permissions (read from X, write to Y, not Z)
• Audit logs for AI-generated changes

The shift: Engineering didn't build the apps. Engineering built the conditions under which apps could be built safely.

Hosted on Acast. See acast.com/privacy for more information.

Is the SaaS model breaking? When AI agents do the work, per-seat pricing stops making sense. Here is what is changing and how founders, CFOs and buyers should prepare.

So, quick context before we dive in. A couple of weeks ago I published a piece on my blog about how AI agents are quietly breaking the SaaS pricing model. And honestly? I didn't expect what happened next. The post just… took off. My inbox has been wild. CFOs, founders, a few VCs, even a couple of procurement leads who I'm pretty sure have never emailed anyone voluntarily in their lives. All asking the same kinds of questions.

Hosted on Acast. See acast.com/privacy for more information.

Previous Gemma licensing:

• Custom "Gemma Terms of Use"
• Usage-policy provisions
• Constraints on commercial deployment

Apache 2.0:

• Fine-tune for commercial use ✓
• Redistribute fine-tuned variants ✓
• Embed in commercial products ✓
• No ongoing license obligations ✓

What's new:

• Full conversational AI on phones, offline
• No data leaving device
• No API costs
• No connectivity requirements

Use cases:

• Healthcare apps (privacy)
• Education (offline areas)
• Finance (data sovereignty)
• Any privacy-sensitive application

The shift:

• European regulators increasingly uncomfortable with US-hosted APIs
• GDPR requires either locked regions or self-hosted
• Gemma 4 + Apache 2.0 = viable self-hosted option
• Regulated industries now unblocked

For Western organizations considering Chinese open models:

1. Training data provenance — Can you verify?
2. Embedded refusals/biases — Different content policies
3. Export-control compliance — Check with legal
4. Strategic precedent — Building on competitor infrastructure

Not disqualifying, but requires conscious decision

Hosted on Acast. See acast.com/privacy for more information.