davidhunter

Using your analogy, imagine it's the year 2026. Two armies are fighting. One uses letter to communicate. One uses phones. Which army do you want to fight in?

This is an obviously poor policy.

I cycle 60 mins per day along the tow path in London on my Brompton, put it under my desk in the office, and then get the train back in the evening. No issues handling that distance.

If you don’t understand then you should invest some time learning microeconomics, marketing, and moats. Principles from (at least) those 3 areas are involved here.

To give 3 examples:

1. The marginal value of these products is in the mind of the individual buyer. No individual is buying both the AirPods Max 2 AND the MacBook Neo for personal use. You can’t compare marginal value across two different individuals.

2. The MacBook Neo has a different set of substitutable goods vs the AirPods Max 2. This affects margin. AirPods Max 2 buyers are likely heavily bought into the Apple ecosystem already.

3. With the Neo, Apple are in some sense subsidising entry into the Apple Ecosystem and ‘getting them young’. Wouldn’t surprise me if there’s zero or negative margin. With the AirPods Max 2 they are exploiting people who are already bought into the ecosystem. Margins will be high.

There's the Tony Hoare Room [1] in the Robert Hooke Building. We held our Reinforcement Learning reading group there.

[1] https://www.cs.ox.ac.uk/people/jennifer.watson/tonyhoare.htm...

Seems like the role of the human operator in the age of AI is to be the entity they can throw in jail if the machine fails (e.g. driver, pilot)

"No, I am not a horse."

Horse rumours denied.

Top of the market

Optimal | London, UK | ONSITE

Simulation and Control Engineer: Up to £150k + 2% depending on experience.

Full-stack Software Engineer (Python, React): Up to £150k + 2% depending on experience.

Reach out directly to me (founder): david@optimal.ag

Optimal is building agents to control the world’s most critical infrastructure - from factories, to datacenters, to farms.

We are backed by the Director of AI Research at Google DeepMind as well as early VC investors in SpaceX, Anduril, and Palantir.

We have built the world’s most advanced climate control system for high-tech greenhouses and have customers in North America and Europe.

https://www.optimal.ag

I’d suggest reading about competitive moats and where they come from. The ability to replicate another’s software does not destroy their moat.

Optimal | London, UK | ONSITE

Backend Software Engineer (Python): Up to £150k + 2% depending on experience.

Optimal builds AI agents to control the world’s critical infrastructure - from factories, to datacenters, to farms.

We are backed by the Director of AI Research at Google DeepMind as well as early VC investors in SpaceX, Anduril, and Palantir.

We have built the world’s most advanced AI control system for high-tech greenhouses and have just signed out first customer contracts in North America and Europe having proven the performance of our AI across 3 seasons in our own demonstration greenhouse.

david@optimal.ag

https://www.optimal.ag

Anyone else think “vibe coding” is a poor choice of name. Too ambiguous.

Friend of mine suggested “apping”.

I ‘apped’ this in 2 hours vs I ‘vibe coded’ this in 2 hours.

Optimal | London, UK | ONSITE

AI Simulation and Control Engineer: Up to £150k + 2% depending on experience (https://wellfound.com/l/2AUs7A)

Backend Software Engineer (Python): Up to £150k + 2% depending on experience (reach out direct to david@optimal.ag)

Optimal builds AI agents to control the world’s critical infrastructure - from factories, to datacenters, to farms.

We are backed by the Director of AI Research at Google DeepMind as well as early VC investors in SpaceX, Anduril, and Palantir.

We have built the world’s most advanced AI control system for high-tech greenhouses and have just signed out first customer contracts in North America and Europe having proven the performance of our AI across 3 seasons in our own demonstration greenhouse.

david@optimal.ag

https://www.optimal.ag

Bill Gates also lost money on greenhouses [0][1]

[0] https://www.reuters.com/technology/bill-gates-green-tech-fun...

[1] https://techcrunch.com/2022/11/03/iron-ox-lays-off-50-amount...

Optimal | London, UK | ONSITE

AI Simulation and Control Engineer: Up to £150k + 2% depending on experience (https://wellfound.com/l/2AUs7A)

Full-Stack Software Engineer (Python): Up to £150k + 2% depending on experience (reach out direct to david@optimal.ag)

Optimal builds AI agents to control the world’s critical infrastructure - from factories, to datacenters, to farms.

We are backed by the Director of AI Research at Google DeepMind as well as early VC investors in SpaceX, Anduril, and Palantir.

We have built the world’s most advanced AI control system for high-tech greenhouses and have just signed out first customer contracts in North America and Europe having proven the performance of our AI across 3 seasons in our own demonstration greenhouse.

david@optimal.ag

https://www.optimal.ag

Optimal | London, UK | ONSITE

AI Simulation and Control Engineer: Up to £150k + 2% depending on experience (https://wellfound.com/l/2AUs7A)

Optimal builds AI agents to control the world’s critical infrastructure - from factories, to datacenters, to farms.

We are backed by the Director of AI Research at Google DeepMind as well as early VC investors in SpaceX, Anduril, and Palantir.

We have built the world’s most advanced AI control system for high-tech greenhouses and have just signed out first customer contracts in North America and Europe having proven the performance of our AI across 3 seasons in our own demonstration greenhouse.

david@optimal.ag

https://www.optimal.ag

That's a nice idea. Thanks for sharing.

I have a single sheet per account (current accounts, share accounts etc). I download csvs and append to the relevant sheet - usually once per month.

In each sheet I've added a column called 'tag'. And I just tag anything that I want to keep track of - which is a small percentage of transactions. Then I can filter transactions by that tag.

Whilst it was a nice idea in theory to book every transaction to an account in a chart of accounts, I found that I very rarely looked at the PnL. And so it didn't justify the time involved in booking each transaction.

Has anyone else gone on the following journey:

1. Use excel

2. See ledger/hledger. Think this must be 'the way'. Go all in.

3. Constantly wrestle with ledger/hledger because you only do your accounting once per month/quarter which is not enough frequency to really grok it.

4. Use excel with a new sense of calm that you're not missing out on something better

AI-generated content marketing has finally got to the front page of HN

The answer my friend, is rovin’ in the wind

Skin in the game for evangelists

Did anyone else check each item manually one-by-one to ensure that the system was accurate?

Investigative journalism report into Boeing by Al Jazeera in 2014:

https://www.aljazeera.com/program/investigations/2014/7/20/t...

The people who are said it wouldn't work lost credibility as these startups raised rounds at progressively higher valuations. After all, money talks, so clearly the naysayers were missing something...

After the later rounds closed at insane valuations, the early VC investors in infarm were seen as legends. This created jealously amongst other VC investors who felt they had 'missed out' which resulted in more money being invested into infarm and other vertical farming companies e.g. Bowery/Plenty.

Decisions to invest were made on FOMO, not a first-principles analysis of the farming technique which shows very clearly that vertical farming is unfeasible.

Would be interested to chat and hear your views on the viability of the greenhouse industry in the US if you're up for it? david [at] optimal.ag

https://en.wikipedia.org/wiki/Tilt_(poker)

Yes. I am like you in that I cannot actually see anything visual in my minds eye but I can still 'visualise' it. For example, I can rotate a die in my minds eye without actually seeing it. It's hard to explain.

It was a revelation when I found out that most people can actually see things visually in their minds eye.

A friend of mine can actually place imagined objects into their field of view, like AR.

Yes, predicting the stock market does involve predicting what other market participants (and hedge fund bros) are going to do.

But, to predict what other market participants (and hedge fund bros) are going to do, you need to predict any world dynamic that will have an eventual effect on stock pricing.

The most successful quantitative funds (RenTech, 2sigma etc) consistently make billions of dollars in cash each year because they have collected the data sets that allow them to do this better than others. But they are still a long long long way off from having a true world model.

Software products have near-zero marginal cost of production and distribution. Physical products have significant marginal cost of production and distribution.

Monopoly dynamics for ubiquitous products comes in large degree from the economies of scale across production and distribution. This allows larger companies to supply the same thing for less money. Even if you eliminated branding, you could not produce and supply a coca-cola can at the same price as coca-cola can :)

With software, you cannot get the same cost advantage through scale.

I hope Tyler Cowen can ask Marc Andreessen how AI works so that we can all learn something from the master