◈ X-Research
← Latest Updates
Key developments
+3
events added
66 total
Events added this run
OpenAI GPT-5.5: Agent Runtime Framing ServiceNow SaaS Referendum (-17.7%) DeepSeek V4 Flash + Pro
Category Events →
Post ideas
ServiceNow paradox — strong AI results, punished anyway
S
Seva Ustinov
@sevaustinov

ServiceNow's Now Assist grew 130% YoY. Stock fell 18%. The market doesn't believe the AI story even when the numbers are real. The new SaaS question isn't 'do you have AI?' It's 'does AI drive net new revenue — or just cannibalize your seats?'

SaaS AI referendum market signal 243/280 ✓
AutoResearch + the goal function is everything THREAD
S
Seva Ustinov @sevaustinov 1/3

Karpathy's AutoResearch ran 700 experiments in 2 days and found 20 optimizations that actually worked. No sleep. No status quo bias. No missed windows. The bottleneck in ML research was never compute — it was human availability.

228/280 ✓
S
Seva Ustinov @sevaustinov 2/3

The same thing is true in performance marketing. We already knew what to test. The bottleneck was how fast we could run and analyze. An agent running 24/7 doesn't change what's possible — it changes how fast you get there.

222/280 ✓
S
Seva Ustinov @sevaustinov 3/3

Greg Isenberg framed it best: 'give it a goal like lower customer acquisition cost.' The goal function IS the skill now. If your objective is clear, an agent can run 100 experiments overnight to find the path.

209/280 ✓
extends Seva's 'humans are the bottleneck' frame to autonomous experimentation
DeepSeek V4 makes model choice a procurement question
S
Seva Ustinov
@sevaustinov

DeepSeek V4 Flash: $0.14/M input tokens. Frontier models: $5-30/M. That's 35-200x cheaper with near-comparable performance. 'Which model does this product run on?' just became a real procurement question every AI vendor will face in 2026.

Open-source cost floor as buyer power shift 238/280 ✓
Comment opportunities
S
Sam Altman @sama · Apr 23

We believe in iterative deployment; although GPT-5.5 is already a smart model, we expect rapid improvements. Iterative deployment is a big part of our safety strategy; we believe the world will be best equipped to win at the team sport of AI resilience this way.

Reply · iterative deployment mirrors buyer adoption curve — connects to Seva's yolo mode trust curve
S
Seva Ustinov
@sevaustinov

Iterative deployment is the right call for both sides. Buyers adopt agents the same way: 10% of workflows first, check results, expand. The vendors that win will be the ones whose iteration cadence matches the buyer's trust curve — not the model release pace.

259/280 ✓
T
Tomasz Tunguz @ttunguz · Apr 22
View on X ↗

The most important market in AI isn't chatbots or search. It's coding. Cursor is the fastest-growing developer tool in history, at $2 billion in annualized revenue.

Reply · extends Seva's 'vibe coding moves complexity up one level, not eliminates it' frame
S
Seva Ustinov
@sevaustinov

Cursor at $2B ARR is proof that vibe coding doesn't kill developer tools — it moves complexity up one level. Simple things get built in-house. The orchestration, integration, reliability layer is where the real product lives. Same story is playing out in every vertical.

270/280 ✓