lifanfzeng - Posthaven

AI Companies

In November 2026 three factors matter:

1. Models

2. Data

3. Harness

Earlier this year models seemed like the only thing that mattered. But now it seems the vast majority of usecases will involve commoditized models, both in terms of cost as well as access. On some tasks model quality still matter, first and foremost coding.

Andrew Ng was right that data is the new oil. But like the world itself, the internet has shifted to be more closed-off. It's unclear what businesses will be able to continually generate useful proprietary data.

Harnesses face pressures from model providers both in terms of cost (hence inference/inhouse training booms) as well as direct competition (why use a wrapper company when I can use Claude for X?). Key question is: what % of business processes can be captured in closed-form?

(GTM is yet another completely separate factor.)

The reviewer in the arena

Let's bucket workers into three categories:

1. ICs

2. (Front-line) Managers

3. Execs+

and then two dimensions: whether youre judged on inputs/outputs, and whether youre directly responsible for something

ICs: directly responsible, but judged in inputs

Managers: directly responsible, judged on outputs

Execs: not directly responsible, judged on outputs

By reducing friction to generating work, the owners of quality (i.e. managers) are pressured on both sides. From ICs, who aren't judged based on the long-term results, and execs/non-R&D, who don't have to be directly responsible for quality.

Some recent observations

U-shaped caring

When people aren't doing well economically, they care a lot about what others think about them. This makes sense because you might need ask for help from others/give to others if they ask. While on the come-up, you care little about what others think, because you're focusing on the economics. But once you've made it, eventually revert back to phase one again. Less out of material necessity, and more for psychological necessity. Possibly there's some element of aiming for the wrong thing

The biggest problem

Gerald Weinberg noted that companies with a lot of problems were healthier. This is because whenever you solve your #1 problem, #2 gets a promo. So good companies that solve problems end up with a lot of new ones. Compare this to having a single problem, e.g. we're running out of money. Here's an extension to the emotional realm: people feel their #1 problem with the same level of intensity, irregardless of the actual seriousness of the problem. Your mileage may vary.

Sources of humor

1. Increasing common knowledge

2. Lowering status

3. Doing something super weird, especially physically

Intermediate dance tips

1. The easiest way to get better is also the funnest, which is to listen to the music

2. Don't coach during the dance

3. If you're offtime, stop and restart

4. Start understanding how different people prefer different pressure

5. Prevent personal hygiene issues

6. New moves will soon stop being the primary measure of progress

7. Wear earplugs