10 yrs in code 🔥 | Frontend • Backend • Productivity | Helping devs ship faster & think sharper 🚀
Joined October 2016
- Tweets 1,911
- Following 351
- Followers 117
- Likes 171
36 Photos and videos
the AI flywheel for personal growth:
learn something new → use AI to apply it → get results → learn from results → use AI to learn faster → repeat.
this flywheel turns faster with AI than without.
because AI compresses the "learn → apply → learn" cycle from weeks to days.
the person spinning this flywheel for 12 months will have 5 years of growth.
the compounding of AI-assisted learning is the ultimate career hack.
start spinning.
1
8
the honest state of AI in mid-2026:
what's working: coding assistants, writing tools, search, data analysis, customer support.
what's not working yet: fully autonomous agents, reliable reasoning over complex problems, video understanding at scale.
what's improving fast: multimodal understanding, agent capabilities, model efficiency.
the gap between hype and reality has narrowed significantly.
we're entering the pragmatic era. and it's the best era yet.
14
the honest state of AI in mid-2026:
what's working: coding assistants, writing tools, search, data analysis, customer support.
what's not working yet: fully autonomous agents, reliable reasoning over complex problems, video understanding at scale.
what's improving fast: multimodal understanding, agent capabilities, model efficiency.
the gap between hype and reality has narrowed significantly.
we're entering the pragmatic era. and it's the best era yet.
14
the AI tool I recommend most often:
not ChatGPT. not Claude. not Cursor.
it's PERPLEXITY.
because everyone searches. every day. multiple times.
replacing Google with Perplexity is the single lowest-friction, highest-impact AI adoption.
no workflow change needed. just search differently.
and the results are dramatically better for any question that needs a real answer (not just links).
start with Perplexity. add other tools later.
the gateway drug to AI productivity.
19
the AI tool I recommend most often:
not ChatGPT. not Claude. not Cursor.
it's PERPLEXITY.
because everyone searches. every day. multiple times.
replacing Google with Perplexity is the single lowest-friction, highest-impact AI adoption.
no workflow change needed. just search differently.
and the results are dramatically better for any question that needs a real answer (not just links).
start with Perplexity. add other tools later.
the gateway drug to AI productivity.
1
18
the simple truth about AI and your career:
you have 3 options:
1. ignore AI and hope it doesn't affect you (it will)
2. resist AI and fight the change (you'll lose)
3. embrace AI and ride the wave (you'll thrive)
option 3 is the only rational choice.
and it starts with opening ChatGPT or Claude and asking: "how can you help me with my job?"
that one question is the beginning of everything.
ask it today.
3
the AI economy is creating 3 new classes:
1. AI capital owners: own the models, infrastructure, data → extreme wealth
2. AI operators: use AI tools expertly → comfortable wealth
3. AI displaced: replaced by AI with no transition → economic pain
the policy question: how do we maximize class 2 and minimize class 3?
retraining. education. transition support. new social contracts.
the technology creates the classes. policy determines their sizes.
5
the AI advice that changed everything for me:
"stop using AI to do things faster. start using AI to do things you couldn't do before."
faster is 1x improvement.
new capabilities are 10x improvement.
faster email writing: nice.
AI-powered customer insights you never had before: game-changing.
the real value of AI isn't speed. it's CAPABILITY.
what can you do now that was impossible before? that's where the magic is.
3
a reflection at tweet 700:
I've written about AI every day for months. and I'm more convinced than ever:
AI is the most important technology of our lifetime.
not because of what it can do today. because of what it ENABLES tomorrow.
every AI capability today is a platform for capabilities tomorrow.
the compounding hasn't even started.
we're at the very beginning. and the beginning is already incredible.
imagine what's next.
6
the AI productivity paradox:
the tools get better every month.
but most people's productivity doesn't improve every month.
why? because the bottleneck isn't the tool. it's the WORKFLOW.
better tools same workflow = marginal improvement.
better tools better workflow = transformative improvement.
the people seeing 2-5x productivity gains aren't using better tools.
they redesigned their workflows around AI.
the workflow is the unlock. not the tool.
1
1
10
the biggest AI opportunity in 2026: industry-specific AI tools.
general AI (ChatGPT, Claude) is saturated. everyone uses them.
but industry-specific AI? wide open.
AI for dentists. AI for real estate agents. AI for restaurants. AI for law firms.
the general AI companies can't go deep enough into each vertical.
that's YOUR opportunity. go deep. build for one industry. own it.
the riches are in the niches. even in AI.
15
the AI adoption curve looks like this:
2023: innovators (5% of companies)
2024: early adopters (15% of companies)
2025: early majority (35% of companies)
2026: late majority (35% of companies) ← we are here
2027: laggards (10% of companies)
if you're adopting AI now, you're in the late majority.
not early. not late. right on time for the mainstream.
the early adopters took the risks. you get the proven playbooks.
that's actually an advantage. learn from their mistakes.
5
Jun 13
the most important AI metric for 2026-2027: revenue per employee.
traditional company: $200K revenue per employee.
AI-native company: $500K-1M revenue per employee.
the gap is 2.5-5x.
this metric will determine which companies survive and which don't.
because higher revenue per employee means: better margins, faster scaling, more competitive pricing.
the companies not tracking this metric are the ones it will hurt most.
3
Jun 13
AI replaced my least favorite part of consulting: the repetitive analysis.
client data analysis that used to take 2 days: now 2 hours.
market research that used to take a week: now 1 day.
report writing that used to take 3 days: now 1 day.
total time saved per project: 60%.
what I do with the saved time: think deeper. talk to more stakeholders. deliver better recommendations.
AI handles the analysis. I handle the insight.
that division of labor is the future of consulting.
11
Jun 13
the AI tipping point for businesses:
when 3 departments use AI daily → the company hits critical mass.
before critical mass: AI is an experiment.
after critical mass: AI is the culture.
the tipping point isn't a technology decision. it's a cultural shift.
and cultural shifts happen suddenly. one month nobody uses AI. the next month everyone does.
the companies approaching critical mass right now will be unrecognizable in 12 months.
3
Jun 13
why the smartest people I know are ALL using AI:
not because they're lazy.
not because they can't do the work.
because they understand leverage.
the smartest people ALWAYS adopt leverage tools first:
→ spreadsheets in the 80s
→ internet in the 90s
→ smartphones in the 2000s
→ AI in the 2020s
each tool multiplied their existing ability.
AI is the ultimate leverage tool. and smart people recognize leverage.
6
Jun 13
the AI productivity stack ranked by ROI:
#1: AI coding assistant (Cursor/Copilot) → 10-20x ROI for developers
#2: AI writing assistant (Claude/ChatGPT) → 5-10x ROI for everyone
#3: AI search (Perplexity) → 3-5x ROI for researchers
#4: AI meeting tools → 2-3x ROI for managers
#5: AI design tools (Midjourney) → 2-5x ROI for creators
the pattern: the closer AI is to your core work, the higher the ROI.
start with the tool closest to your daily work.
35
Jun 13
AI hot takes that are actually cold reality:
→ most AI startups will fail (just like most startups)
→ AI won't cause mass unemployment (but will cause mass displacement)
→ open-source AI will rival proprietary (already happening)
→ AI regulation will lag behind innovation (always has)
→ the best AI use cases are boring ones (enterprise, not consumer)
these aren't hot takes. they're observable patterns.
the signal is in the boring truths, not the exciting predictions.
6
Jun 13
the most important thing I've learned about AI in 3 years:
it's not about the AI. it's about the HUMAN using it.
the same AI tool in different hands produces wildly different results.
the variable isn't the technology. it's the user.
expertise AI = excellence.
ignorance AI = expensive mistakes.
invest in yourself first. then add AI.
the best AI in the world can't fix unclear thinking.
3
Jun 13
the AI landscape is splitting into two worlds:
world 1: big tech AI (OpenAI, Google, Anthropic, Meta)
→ massive models, massive compute, massive investment.
world 2: small/specialized AI (niche startups, fine-tuned models)
→ focused models, efficient compute, bootstrapped.
both worlds are thriving. for different reasons.
big AI wins on breadth. small AI wins on depth.
the future isn't one world winning. it's both coexisting.
choose your world based on your problem.
15