Possible Futures with AI — sl.md

sl.md

Possible Futures with AI

an semi-astrological essay for IT people and not only

Nikolai Ryzhikov

---

What will happen with your job and company in the next 2-5 years?

---

Electronics middle market collapse

I see parallels in the history of the IT and Electronics

in the 1920s-1960s, Electronics was a distributed cottage industry with tens of thousands of small and mid-sized businesses.

Then, the integrated circuit was invented in 1958. This made the production of electronics much more efficient and cheaper.

As a result, the middle market of electronics collapsed.

---

Let's look what happened in Electronics

1920s-1960s:

• ~150 TV manufacturers in the US alone (early 1950s)
• ~20,000 electronics repair shops
• ~1,500 component distributors
• 1.9M workers in electronics manufacturing
• 24,000 EE graduates/year at peak (1985)

---

After the integrated circuit:

jobs lost: 1.9M → 1.0M (47% decrease) global fab share: 42% (1980) → 12% (2020) Fab cost: $4M (1970) → $20B (2020s)

---

What survived?

Niches too small or too demanding for mass production:

• Medical devices, aerospace, defense
• Fabless design companies (Qualcomm, Nvidia, AMD)
• Contract manufacturers for custom small runs

---

What's happend to the jobs?

No need in routine work! Still need R&D, but not in the same quantity.

R&D survived: 21-28% of revenue

---

Let's interpolate this to the IT middle market

How many IT jobs are now?

50-60M

---

How much of these jobs are routine?

• 10-20% right now
• 40-50% in next 10 years (probably underestimating)

---

What is software factory?

We already sow platforms wave one - clouds, crm, erp, ecom etc. ~30% is just configuration and customization of platforms. AI already automates this work.

Formula: [Platform] + [AI]

---

What we can learn from electronics?

Production of something like TV will be monopolized by a few companies.

---

Will all current platforms survive?

History of IBM: - probably not - there is a quick opportunity to build a new platform for post-AI era.

Local LLMs vs commercial models

OpenAI, Anthropic took crazy credits for their models. They think that they will be the only ones to provide models and power of this models will scale.

But let's imagine it's not! It will rich the logical platoe. And we will have open weights models which we can run locally.

Bright side:

When electronics collapsed, computers and programming were born

• creating new industries.

What is computers/programming industry after AI disruption?

---