Engaging with genAI Beyond the Hype

by Erin Casali

Embed

Start on current slide

Slide 1

Slide 1 text

Engaging with genAI Beyond the Hype Erin Casali

Slide 2

Slide 2 text

MSc + Patent Virtual Assistant Design AI Features Advisor AI Startups Product & Design AI Strategy Make & Learn AI Personal mini-apps & experiments SMARTER, SIMPLER, SOCIAL

Slide 3

Slide 3 text

Big picture

Slide 4

Slide 4 text

Societal impact Important, but not today’s topic Privacy Jobs Copyright Psychosis Biases Climate Misinformation

Slide 5

Slide 5 text

Seen as hater by lovers Seen as lover by haters

Slide 6

Slide 6 text

Diminishing core tech returns. Shift: harnesses, tool use, UI. Signal MMLU-Pro (2026 Feb 14) Graph by Alex Ellman (2026 Feb 14)

Slide 7

Slide 7 text

Signal Compute costs going down. Arti fi cialAnalysis.ai/trends (2026 Feb 14)

Slide 8

Slide 8 text

Signal Open source models run local. LMStudio Kimi K2.5 GLM 4.7 Qwen3 Gemma3 DeepSeek R1

Slide 9

Slide 9 text

Fears

Slide 10

Slide 10 text

I worry I won’t like my job anymore

Slide 11

Slide 11 text

Counter Recenter on what you enjoy Where can genAI support that?

Slide 12

Slide 12 text

SKILLS REFRAMING Vibe  coding Vibe  engineering Write code with LLMs without reviewing the code written Build software using production-level engineering practices

Slide 13

Slide 13 text

I worry I will be replaced

Slide 14

Slide 14 text

Counter Do not believe the hype What can genAI actually do?

Slide 15

Slide 15 text

AI hallucinations spring from their fundamental construction — Dilek Hakkani-Tür (CS Professor / Principal Scientist Alexa AI) “ LLMs will inevitably hallucinate if used as general problem solvers — Z. Xu, S. Jain, M. Kankanhalli (2025) Hallucination is Inevitable “ 7~10 % about 1 out of 12 HALLUCINATION RATE Calibration—and, hence, errors—is a natural consequence — A. T. Kalai, O. Nachum, S. S. Vempala, E. Zhang (OpenAI) “ Vectara, 5 Feb 2026 7.8 % Gemini 2.5 Flash 10.8 % ChatGPT 5.2 High 10.9 % Claude Opus 4.5 7.0 % Gemini 2.5 Pro

Slide 16

Slide 16 text

I worry I can’t keep up with change

Slide 17

Slide 17 text

Counter Do not get sucked in Can you skim to fi nd patterns?

Slide 18

Slide 18 text

Most AI news   is PR Find   good sources

Slide 19

Slide 19 text

Adoption

Slide 20

Slide 20 text

Di ff usion of Innovation INNOVATORS EARLY   ADOPTERS EARLY   MAJORITY LAGGARDS LATE   MAJORITY E. Rogers (1962)

Slide 21

Slide 21 text

Adoption of AI INNOVATORS EARLY   ADOPTERS EARLY   MAJORITY LAGGARDS LATE   MAJORITY Microsoft (2026) “Global AI Adoption in 2025”

Slide 22

Slide 22 text

Secondary Waves of Change INNOVATORS EARLY   ADOPTERS EARLY   MAJORITY LAGGARDS LATE   MAJORITY

Slide 23

Slide 23 text

Positioning Strategies INNOVATORS EARLY   ADOPTERS EARLY   MAJORITY LAGGARDS LATE   MAJORITY Pioneers Adopters

Slide 24

Slide 24 text

Pioneers Time to dedicate Dead ends Fun from exploring

Slide 25

Slide 25 text

Adopters Don’t worry Wait for case studies Wait for features Befriend pioneers

Slide 26

Slide 26 text

News

Slide 27

Slide 27 text

To navigate AI news ask yourself WHAT?!

Slide 28

Slide 28 text

AUTOMATION & DECISION MAKING REASONING & DEDUCTION PREDICTION & FORECASTING GENERATION & TRANSFORMATION PATTERN RECOGNITION PERSONALIZATION WHAT?! Winnow Rich open natural dialogue Manipulate & summarize text Generate content Generate synthetic data Extract structured data Transcribe speech Translation Classi fi cation & tagging Extract text from images Identify objects in images Trend analysis (past) Forecasting (future) Anomaly detection Recommendations Adaptive interfaces Optimization Planning & sequencing Logical inference AI Capabilities Map

Slide 29

Slide 29 text

WHAT?! Winnow AUTOMATION & DECISION MAKING REASONING & DEDUCTION PREDICTION & FORECASTING GENERATION & TRANSFORMATION PATTERN RECOGNITION PERSONALIZATION LLM LLM, ML Transformer Networks Transformer Networks, BERT CNN, Multi-modal LLM ML ML LLM, ML Expert Systems LLM ML, Symbolic Reasoning AI Capabilities Map: Underlying Tech

Slide 30

Slide 30 text

WHAT?! Winnow Action Swap “AI” mentions with the speci fi c tech in your brain.

Slide 31

Slide 31 text

WHAT?! Herds Doomers Boomers “AI is inevitable, and will save the world (we must be fi rst)” “AI is inevitable, and will destroy the world (we must be fi rst)” Karen Hao (2025)

Slide 32

Slide 32 text

WHAT?! Herds Users Destroyers “AI is horrible, we need to stop it at any cost” “AI is here, I just want to use it if it’s ok with y’all”

Slide 33

Slide 33 text

WHAT?! Herds Identify which side is speaking in the news you are reading. Action

Slide 34

Slide 34 text

WHAT?! Adoption Look for patterns and trends

Slide 35

Slide 35 text

WHAT?! Adoption Rule of three: once you see the same concept in the news thrice, check it. Action

Slide 36

Slide 36 text

WHAT?! Try Do not believe the hype

Slide 37

Slide 37 text

WHAT?! Try Scan for use cases, then give it 10 minutes to try it. Action

Slide 38

Slide 38 text

Wrap

Slide 39

Slide 39 text

Permission to feel If you like, learn and have fun If you do not like, learn and understand

Slide 40

Slide 40 text

Permission to choose Pioneer or Adopter

Slide 41

Slide 41 text

Permission to ignore WHAT?! approach to AI news

Slide 42

Slide 42 text

“ ” Karen Hao ‘Empire of AI’ The name artificial intelligence was a marketing tool from the very beginning

Slide 43

Slide 43 text

Thanks. IntenseMinimalism.com — READ MORE & FOLLOW ON —