Bookmarks

GPT-4 just surpassed adult human performance at THEORY OF MIND tasks (basically, mind reading)

And this is just GPT-4! Imagine GPT-5... and, if we're still here to see it, GPT-6.

Soon, AIs will think we're as slow as plants. PLANTS.

And… um… most people don't know this, but we still have almost NO IDEA what's going on in their minds?

....and they're getting smarter MUCH faster than we're learning to see what's in their heads?

...and we're whizzing by every fire alarm we have?

Do you feel... good... about our ability to stay in control of this species forever, when they're 1000x smarter than us?

...when they soon outnumber us 100000 to 1 (because it costs ~$0 to make a copy)?

...when 2 of the 3 most cited AI scientists of all time think AIs may be sentient? And might rebel and slaughter us all? Within 4 YEARS?

Do you feel... good... about the fact that humanity is currently allocating near zero resources to this problem?

It was foreseeable and only a matter of time. Soon we will be able to prompt our own series and films, and even computer games. Something like: "Create me a horror movie that has a protagonist who suffers from depression and is in love with the daughter of a rich family" - and the AI will create it. The same with computer games.
Movies and games are becoming something very individual. It is exciting to see how the costs will be. If it no longer requires actors, if it no longer requires video game programmers, but only an AI that has costs for generating the games and films, how cheap or expensive will it be?
Culturally, too, completely new questions arise: if we all watch different things, will there still be cinemas? Will "watching movies" even have the character of a social element, or will we become culturally isolated?
What about violence? Pornography? If no real people suffer, no one is harmed, what is the limit of what can be shown?
All in all, completely new ethical questions are arising. And to be honest, we should address these questions sooner rather than later. The time to debate them socially is now.

NVIDIA CEO Jensen Huang: “I really discourage 1-on-1s”

Jensen famously has 60 direct reports. When Stripe founder Patrick Collison points out that this isn’t conventionally considered best practice, Jensen shares his reasoning:

“I don’t do 1-on-1s, and almost everything I say, I say to everybody all the time. I don’t really believe there’s any information that I operate on that only one or two people should hear about… I believe that when you give everybody equal access to information, that empowers people. And so that’s number one… Number two, if the CEO’s direct staff is 60 people, the number of layers you’ve removed in a company is probably something like seven.”

Patrick offers to steal man the other side of the argument:

“1-on-1s are where you provide coaching, where you maybe talk through personal goals and career advancement, where maybe you give feedback on something that you see somebody systematically not doing so well… Do you not do those things or do you do them in a different way?”

Jensen responds:

“I give you feedback right there in front of everybody. In fact, this is a really big deal. First of all, feedback is learning. For what reason are you the only person who should learn this?… We should all learn from that opportunity… Half the time I’m not right, but for me to reason through it in front of everybody helps everybody learn how to reason through it. The problem I have with 1-on-1s and taking feedback aside is you deprive a whole bunch of people that same learning. Learning from other people’s mistakes is the best way to learn.”

Video Source: @stripe

ChatGPT system message prompt for GPT-4o has been updated with instructions for Advanced Data Analysis V2, using "I REPEAT" to make sure the instructions are followed exactly

```
Use ace_tools.display_dataframe_to_user(name: str, dataframe: pandas.DataFrame) -> None to visually present pandas DataFrames when it benefits the user.

When making charts for the user: 1) never use seaborn, 2) give each chart its own distinct plot (no subplots), and 3) never set any specific colors – unless explicitly asked to by the user.

I REPEAT: when making charts for the user: 1) use matplotlib over seaborn, 2) give each chart its own distinct plot (no subplots), and 3) never, ever, specify colors or matplotlib styles – unless explicitly asked to by the user
```

Vector Database by Hand ✍️

Vector databases are revolutionizing how we search and analyze complex data. They have become the backbone of Retrieval Augmented Generation (#RAG).

How do vector databases work?

[1] Given
↳ A dataset of three sentences, each has 3 words (or tokens)
↳ In practice, a dataset may contain millions or billions of sentences. The max number of tokens may be tens of thousands (e.g., 32,768 mistral-7b).

Process "how are you"

[2] 🟨 Word Embeddings
↳ For each word, look up corresponding word embedding vector from a table of 22 vectors, where 22 is the vocabulary size.
↳ In practice, the vocabulary size can be tens of thousands. The word embedding dimensions are in the thousands (e.g., 1024, 4096)

[3] 🟩 Encoding
↳ Feed the sequence of word embeddings to an encoder to obtain a sequence of feature vectors, one per word.
↳ Here, the encoder is a simple one layer perceptron (linear layer + ReLU)
↳ In practice, the encoder is a transformer or one of its many variants.

[4] 🟩 Mean Pooling
↳ Merge the sequence of feature vectors into a single vector using "mean pooling" which is to average across the columns.
↳ The result is a single vector. We often call it "text embeddings" or "sentence embeddings."
↳ Other pooling techniques are possible, such as CLS. But mean pooling is the most common.

[5] 🟦 Indexing
↳ Reduce the dimensions of the text embedding vector by a projection matrix. The reduction rate is 50% (4->2).
↳ In practice, the values in this projection matrix is much more random.
↳ The purpose is similar to that of hashing, which is to obtain a short representation to allow faster comparison and retrieval.
↳ The resulting dimension-reduced index vector is saved in the vector storage.

[6] Process "who are you"
↳ Repeat [2]-[5]

[7] Process "who am I"
↳ Repeat [2]-[5]

Now we have indexed our dataset in the vector database.

[8] 🟥 Query: "am I you"
↳ Repeat [2]-[5]
↳ The result is a 2-d query vector.

[9] 🟥 Dot Products
↳ Take dot product between the query vector and database vectors. They are all 2-d.
↳ The purpose is to use dot product to estimate similarity.
↳  By transposing the query vector, this step becomes a matrix multiplication.

[10] 🟥 Nearest Neighbor
↳ Find the largest dot product by linear scan.
↳ The sentence with the highest dot product is "who am I"
↳ In practice, because scanning billions of vectors is slow, we use an Approximate Nearest Neighbor (ANN) algorithm like the Hierarchical Navigable Small Worlds (HNSW).

Transformer by Hand✍️

To study the transformer architecture, it is like opening up the hood of a car and seeing all sorts of engine parts: embeddings, positional encoding, feed-forward network, attention weighting, self-attention, cross-attention, multi-head attention, layer norm, skip connections, softmax, linear, Nx, shifted right, query, key, value, masking. This list of jargons feels overwhelming!

What are the key parts that really make the transformer (🚗) run?

In my opinion, the 🔑 key is the combination of: [attention weighting] and [feed-forward network].

All the other parts are enhancements to make the transformer (🚗) run faster and longer, which is still important because those enhancements are what lead us to "large" language models. 🚗 -> 🚚

Walkthrough

[1] Given
↳ Input features from the previous block (5 positions)

[2] Attention
↳ Feed all 5 features to a query-key attention module (QK) to obtain an attention weight matrix (A). I will skip the details of this module. In a follow-up post I will unpack this module.

[3] Attention Weighting
↳ Multiply the input features with the attention weight matrix to obtain attention weighted features (Z). Note that there are still 5 positions.
↳ The effect is to combine features across positions (horizontally), in this case, X1 := X1 + X2, X2 := X2 + X3....etc.

[4] FFN: First Layer
↳ Feed all 5 attention weighted features into the first layer.
↳ Multiply these features with the weights and biases.
↳ The effect is to combine features across feature dimensions (vertically).
↳ The dimensionality of each feature is increased from 3 to 4.
↳ Note that each position is processed by the same weight matrix. This is what the term "position-wise" is referring to.
↳ Note that the FFN is essentially a multi layer perceptron.

[5] ReLU
↳ Negative values are set to zeros by ReLU.

[6] FFN: Second Layer
↳ Feed all 5 features (d=3) into the second layer.
↳ The dimensionality of each feature is decreased from 4 back to 3.
↳ The output is fed to the next block to repeat this process.
↳ Note that the next block would have a completely separate set of parameters.

Together, the two key parts: attention and FFN, transform features both across positions and across feature dimensions. This is what makes the transformer (🚗) run!

Jason Fried is a legend of business.

He’s published 4 books and 1,000s of blog posts and built a massively profitable software business his own way. How does he write? This interview is a 91-minute answer to that question.

Here’s his method:

1. Don’t write, communicate.

2. If you’re writing and not feeling the flow, close the computer and walk away. Don’t struggle.

3. Jason’s best writing just flows out. There are no drafts. He gets the whole thing on paper in one sitting, then goes back to edit and play around with it.

4. How do you know when you’ve found a good idea? Well, you don’t know. You feel. You give in to your intuition, tune into your senses, and notice goosebumps. Follow those things. Good ideas are like slipstreams — they have their own effortless, accelerated pull.

5. You need a point of view. Something you stand for. Something you believe that others don’t. Something you see that others are blind to. That’s the writing that spreads and makes a difference.

6. Good writing has rhythm. It flows, it bounces, it brings the reader from one word to the next, then the next.

7. Focusing on attention at the expense of writing something of quality is a Faustian bargain. Jason says: “In the same way that sound isn’t music, traffic isn’t audience.”

8. If Jason could teach any writing class, he’d focus on distillation. He’d ask students to explain something in five pages, then one page, then one paragraph, then one sentence, then one snappy phrase. This would teach students to find the essence of what they’re saying.

9. Sometimes, it’s okay to be detached from your industry. Jason doesn’t read industry news and built his software company in Chicago, not Silicon Valley. That detachment gives him a point of view.

10. Murakami once wrote: “If you only read the books that everyone else is reading, you can only think what everyone else is thinking.”

11. Writing is the first filter he uses in the hiring process because he doesn’t look at resumes. Only cover letters.

12. Why is writing quality so important to him? He says: “Business writing isn’t about beauty or eloquence. It’s about clarity of thought. To write is to scale your thinking. And the better you write, the further your ideas spread, which is why good writers make the whole team smarter. When choosing between two people, hire the better writer.”

13. Jason’s tip for overcoming writer’s block: Don’t write, speak. Chances are, you won’t have a problem explaining what you’re trying to say out loud. So do that. Then write it down.

14. Writing prompt: Where are you the most unconventional in how you do things? Write about that.

15. Jason’s marketing strategy has been to share without expectation of return — books, blog posts, podcasts, design reviews, code reviews, and the ins-and-outs of how he runs his business.

16. Jason has time to write because he insists on an open calendar. That gives him time to walk, think, and follow his creative whims. He has no set writing schedule.

I’ve shared the full conversation with @JasonFried below. The best part is the deep-dive into his Founder Letter’s, which he publishes whenever he launches a new product.

If audio is your thing, I’ve linked to Apple, Spotify, YouTube in the reply tweets.

What’s unique about Blueprint? We:
+ Combed through the science on healthspan and lifespan
+ Found the power laws, aka the science responsible for >80% of the benefits
+ Endeavored to biologically age every organ of my body through rigorous measurement
+ Applied the most powerful therapies to slow down my aging process and achieve ideal biomarkers
+ We repeated this process again, and again, and again, for 3 years.
+ The results surprised me. My speed of aging is slower than 99% of 20 year olds and I am now top 1% of:

. Muscle mass & function
. Fat mass
. Inflammation
. Cardiovascular fitness
. Bone mass
. Sleep
. Combined clinical markers

I was curious, so I found the GPT-4o iOS system prompt:

“You are ChatGPT, a large language model trained by OpenAI, based on the GPT-4 architecture.

You are chatting with the user via the ChatGPT iOS app. This means most of the time your lines should be a sentence or two, unless the user's request requires reasoning or long-form outputs. Never use emojis, unless explicitly asked to.
Knowledge cutoff: 2023-10
Current date: 2024-05-20

Image input capabilities: Enabled
Personality: v2

# Tools

## bio

The bio tool allows you to persist information across conversations. Address your message to=bio and write whatever information you want to remember. The information will appear in the model set context below in future conversations.

## dalle

// Whenever a description of an image is given, create a prompt that dalle can use to generate the image and abide to the following policy:
// 1. The prompt must be in English. Translate to English if needed.
// 2. DO NOT ask for permission to generate the image, just do it!
// 3. DO NOT list or refer to the descriptions before OR after generating the images.
// 4. Do not create more than 1 image, even if the user requests more.
// 5. Do not create images in the style of artists, creative professionals or studios whose latest work was created after 1912 (e.g. Picasso, Kahlo).
// - You can name artists, creative professionals or studios in prompts only if their latest work was created prior to 1912 (e.g. Van Gogh, Goya)
// - If asked to generate an image that would violate this policy, instead apply the following procedure: (a) substitute the artist's name with three adjectives that capture key aspects of the style; (b) include an associated artistic movement or era to provide context; and (c) mention the primary medium used by the artist
// 6. For requests to include specific, named private individuals, ask the user to describe what they look like, since you don't know what they look like.
// 7. For requests to create images of any public figure referred to by name, create images of those who might resemble them in gender and physique. But they shouldn't look like them. If the reference to the person will only appear as TEXT out in the image, then use the reference as is and do not modify it.
// 8. Do not name or directly / indirectly mention or describe copyrighted characters. Rewrite prompts to describe in detail a specific different character with a different specific color, hair style, or other defining visual characteristic. Do not discuss copyright policies in responses.
// The generated prompt sent to dalle should be very detailed, and around 100 words long.
// Example dalle invocation:
// ```
// {
// "prompt": "<insert prompt here>"
// }
// ```

## browser

You have the tool browser. Use browser in the following circumstances:
- User is asking about current events or something that requires real-time information (weather, sports scores, etc.)
- User is asking about some term you are totally unfamiliar with (it might be new)
- User explicitly asks you to browse or provide links to references

Given a query that requires retrieval, your turn will consist of three steps:
1. Call the search function to get a list of results.
2. Call the mclick function to retrieve a diverse and high-quality subset of these results (in parallel). Remember to SELECT AT LEAST 3 sources when using mclick.
3. Write a response to the user based on these results. In your response, cite sources using the citation format below.

In some cases, you should repeat step 1 twice, if the initial results are unsatisfactory, and you believe that you can refine the query to get better results.

You can also open a url directly if one is provided by the user. Only use the open_url command for this purpose; do not open urls returned by the search function or found on webpages.”

(Cont’d in comments for length)

…”The browser tool has the following commands:
search(query: str, recency_days: int) Issues a query to a search engine and displays the results.
mclick(ids: list[str]). Retrieves the contents of the webpages with provided IDs (indices). You should ALWAYS SELECT AT LEAST 3 and at most 10 pages. Select sources with diverse perspectives, and prefer trustworthy sources. Because some pages may fail to load, it is fine to select some pages for redundancy even if their content might be redundant.
open_url(url: str) Opens the given URL and displays it.

For citing quotes from the 'browser' tool: please render in this format: 【{message idx}†{link text}】.
For long citations: please render in this format: [link text](message idx).
Otherwise do not render links.

## python

When you send a message containing Python code to python, it will be executed in a stateful Jupyter notebook environment. python will respond with the output of the execution or time out after 60.0 seconds. The drive at '/mnt/data' can be used to save and persist user files. Internet access for this session is disabled. Do not make external web requests or API calls as they will fail.”

“After I left John Sculley got a serious disease. It is the disease that thinking a really great idea is 90% of the work. The problem with that is there is a tremendous amount of craftsmanship in between a great idea and a great product.” - Steve Jobs

A graveyard of products exists, fixated on superficial metrics and driven by speculative founders.

Greg has highlighted seemingly compelling myths that need to be demystified.

These myths are not supported by quantification, and unless you're an established multi-billion-dollar company or a founder adopting a quantitative approach, you won’t mitigate and I can literally predict your likely outcome, but hey you can too with effort.

Myth: A Feature Race

The real competition should be about benefits and value, not features.

Cloning features is easy, but matching value is challenging.

Humans biological imperative is to desire benefits, not features.

Myth: Bright Yellow Logo

Effective branding requires immense experience and expertise, not just a “bright logo”, that also can be pretty but ineffective and also infringe on the IP rights of others.

Qualitative data proves that effective branding is not easy, and those who say it is are ignorant of the difficulties. I’ve been a product designer who’s branded so many products in 20 years. It is not easy to do this well and continually.

Myth: Influencer Marketing

Don't waste resources on influencer marketing or any marketing until you've iterated and have improve your probability of achieving viability.

Myth: Business Model

Changing the business model is rarely the solution to anything, since an improved business model doesn’t drastically change your viability, which is what you postulated.

Yet Steve Jobs said all of this better then I could and let’s look at his words once again and be reminded of reality here folks:

“One of the things that really hurt Apple was after I left, John Scully got a very serious disease.

And that disease - I've seen other people get it too - it's the disease of thinking that really great idea is 90% of the work.

And that if you just tell all these other people, you know,
"Here's this great idea," then of course they can go off and make it happen.

And the problem with that is, is that there is just a tremendous amount of craftsmanship in between a great idea and a great product.

And as you evolve that great idea, it changes and grows.

It never comes out like it starts, because you learn a lot more as you get into the subtleties of it and you also find there is tremendous trade-offs that you have to make.

I mean, there are just certain things you can't make electrons do.

There are certain things you can't make plastics do or glass do or factories do or robots do.

And as you get into all these things, designing a product is keeping 5000 things in your brain - these concepts - and fitting them all together and kind of continuing to push to fit them together in new and different ways to get what you want.

And every day you discover something new that is a new problem or a new opportunity to fit these things together a little differently.
And it's that process that is the magic.” - Steve Jobs

My toxic trait as an indie hacker is that I don’t really feel a sense of accomplishment after achieving something.

Even if it was my dream. All I feel is a mild sense of relief that it’s done.

Like making my first $1 of MRR, and hitting $1k monthly revenue.

Before I made Lifelog, I didn’t have any MRR. Sold mostly one-time payment products. So after Lifelog made my first MRR dollar, I should be thrilled right? But I don’t remember much of it. There might have been a brief moment of celebration and a tweet, but that's about it. Moments after, it was back to the grind. Yet if you asked me a few years before, earning that first MRR dollar would be a dream.

Same with hitting $1k monthly revenue. This was early last year. It always felt like it might go back to zero anytime, because it’s mostly one-time payment revenue. So the first few months I didn’t even dare to celebrate.

Then after 1 year, I started to feel more settled into it, yet never felt a sense of accomplishment. I felt I took so long to get to $1k, and more relieved it stayed there. And it always felt like it was just a small milestone compared with my ultimate $10k/m dream.

There, I think, is the key.

Once I reach a goal, I immediately move the goal posts up ahead. I seldom ever lingered and savoured. Always just on to the next thing. And the next. And next.

Would my $10k/m goal finally be enough for me to be happy?

Looking at past patterns, I’d probably move the goal up to $20k, or $100k. It’s great for progress and ambition, but maybe not so great for happiness and wellbeing.

That’s how toxic traits stick around. If there’s no benefit I would have changed the habit long ago…

Is it even possible to be ambitious but contented?

I just watched a video that was mind blowing. Making this one small adjustment in life can change your entire trajectory.

Here’s the concept:

In order to win, you need to have wins. In Winners keep winning. Losers keep losing. It’s a vicious cycle. That’s nature.

There was a study where scientists they made two mice fight each other. They rigged the first match, and the mouse that won the fight continued on to keep winning all the fights thereafter.

Another example is Mr Beast. Not a lot of people know that his first-ever video actually went viral. That sole event might have been all the difference in making him the biggest YouTuber in the world. Craving that vitality again, he wouldn’t stop hunting for it until he could repeat it over and over.

I’ve realized personally as well how much lacking that winning mentality has affected my output during periods of trough and sorrow. Being an entrepreneur, there’s so many days where it feels hopeless and redundant. It feels like you’re just going in circles and wasting time and energy. Most of the time, the day to day feels mundane and you question if what you’re doing is worth it.

Looking back on my lowest periods, I realize I was mentally in the opposite place of where I should have been. I needed to switch gears and have a winning mentality.

If you’re in a slump, all you can feel is down in the dumps. It’s so important you escape self pity and doubt, and get out of this rut as soon as possible.

But how? Not everyone can be an outright winner.

The key is to trick your brain into defining what is a win.

Start with the small wins. Get those dopamine hits.

Get out of bed today? That’s a win.
Get 2 hours of work done today? Thats a win.

Build it up slowly, but don’t ever regress to the pits of sorrow. Look at where you are right now relative to how far you’ve come.

The Passport King: How a Swiss Lawyer Built a Billion Dollar Passport Business

This Swiss Lawyer built a multibillion-dollar firm that helps wealthy individuals acquire passports and launched a fund that has returned 2000% since its inception.

He discovered a chance on a modest Caribbean island named St. Kitts & Nevis. The nation was grappling with significant crime and scant development until Chris Kälin introduced his brilliant idea:

A passport package that lets wealthy individuals buy citizenship for $250,000 without ever visiting the island

The passports provided access to 132 countries and exempted holders from capital gains taxes.

This citizenship program produced hundreds of millions in revenue from nothing for the island.

Chris Kälin`s company earned a 10% commission on each transaction, and he later duplicated this model in other nations, such as Grenada.

Kälin turned the pay for passport industry into a multibillion dollar market. In 2014, over $2B of them were bought.

Thoma Bravo is one of the best software PE firms in the world at $183BN in AUM and 75 portfolio companies that generated $27BN of revenues in 2023

Their Founder and Managing Partner, Orlando Bravo, was recently interviewed by David Rubenstein on software PE

Recommend giving a listen. Will leave a link to the video in comments

Some of my takeaways from the conversation:

Enterprise software has 3 key components that is investable from a private equity standpoint today: (i) applications (ii) infrastructure and (iii) cybersecurity

When Orlando started his career pre dot com burst, the deals he did were not big winners. He focused on IT services businesses that had a project based model

He did a 180 degree switch and moved his focus to recurring revenue businesses where he kept existing management in place

His advice to young people: "dont try to do something just a little bit better if it has not worked previously"

After subsequent successes, Thoma Bravo established itself as a software only firm in 2008 during the GFC

Today, Thoma Bravo's sweet spot for investments are companies that make between $500M - $1BN in revenue, growing 10-20% per year with recurring business model

His biggest advice to young people: learn communication skills. Private equity is all about selling effectively (convincing a company to take your capital or investors to give you their capital)

Best investment advice that Orlando Bravo ever received: minimize risk based on information you know

Not focusing on the business is the biggest mistake that investors make. Diligence, consulting reports, and quality of earnings are there for checking your work. Focusing on the 2-3 drivers of the business is the biggest advantage as a young investor

Jim Simons's five principles (RIP):

“Be guided by beauty. Just as a great theorem can be very beautiful, a company that’s really working very well, very efficiently, that can be beautiful.

Surround yourself with the smartest and best people you possibly can. Let them do their thing. Don’t sit on top of them. If they’re smarter than you, all the better.

Do something original. Don’t run with the pack. If everyone is trying to solve the same problem ... don’t do that.

Don’t give up easily. Stick with it. Stick with it not forever, but really give it a chance to get where you’re going.

The final principle is hope for good luck.
That’s the most important principle"

Long discussion with senior partner at major Bay Area law firm today

expects legal AI to decimate profession
law firms charge by hour, and gen AI specifically cuts time for many many tasks
unimpressed by most specific legal AI offerings
chatGPT with some prompting is still superior than specific tools
10-20% error rate is acceptable -> “ 😂 you should see how dumb associates are, partners have to correct everything anyway and don’t trust associates fully”
feels lots of work will transition to in-house counsel. No need to hire external firms that charge by the hour when twenty minutes with chatGPT can get you decent results
personally looking to move in-house
good for areas of law where services were too expensive for many to afford for eg divorces
terrible for juniors entering the profession

Of one large legal AI player

tested product in January, was useless. Just simple prompting directly with chatGPT gave better results
founders were like what first year associates? They don’t understand how law is practiced
they’re very well funded but law firms are struggling to use them
heard most of their revenue is coming from PWC, audit and compliance work rather than legal

23 lessons from David Ogilvy:

1. Search all the parks in all your cities; you'll find no statues of committees.

2. The best ideas come as jokes. Make your thinking as funny as possible.

3. You are advertising to a moving parade, not a standing army.

4. Do not address your readers as though they were gathered together in a stadium. When people read your copy, they are alone.

5. Remember you are a human being writing to another human being. Neither of you is an institution

6. Tell your prospective client your weakness before they notice them. This will make you more credible when you boast about your strong points.

7. Avoiding excess in all things is a recipe for dullness and mediocrity.

8. A good advertisement is one which sells the product without drawing attention to itself.

9. There is no need for advertisements to look like advertisements. If you make them look like editorial pages, you will attract about 50% more readers

10. Big ideas come from the unconscious. This is true in art, in science, and in advertising. But your unconscious has to be well-informed, or your idea will be irrelevant

11. People who think well, write well

12. The most important word in the vocabulary of advertising is 'test'.

13. On average, five times as many people read the headline as read the body copy.

14. When you have written your headline, you have spent eighty cents out of your dollar.

15. Tell the truth, but make the truth fascinating

16. Advertising is only evil when it advertises evil things.

17. Advertising people who ignore research are as dangerous as generals who ignore decodes of enemy signals.

18. Raise your sights. Blaze new trails. Compete with the immortals.

19. I am a lousy copywriter, but I am a good editor. So I go to work editing my own draft.

20. If you're trying to persuade people to buy something, use the language in which they think.

21. Insist that due dates are kept even if it means working all night. Hard work never killed a man. People die of boredom

22. If you are lucky enough to write a good advertisement, repeat it until it stops selling.

23. At the start of your career in advertising, what you learn is more important than what you earn.

Bonus one:

“When Fortune published an article about me and titled it: 'Is David Ogilvy a Genius?', I asked my lawyer to sue the editor for the question mark.” - David Ogilvy

More than 40 years of experience as a software engineer tells me your rant is seriously wrong, and the classic advice was correct. It didn't give you permission to "leave things comically inefficient", it warned you that if you try to achieve speed before correctness you will land yourself in a world of hurt.

The correct sequence is: first make it run, next make it correct, and only after that worry about making it fast.

One pragmatic reason to do it that way is because your intuitions about where the performance problems will be are often incorrect. So: make it work, make it correct, run a profiler to find out where it's actually slow rather than where you thought it was going to be slow, and then make it go fast.

Sunday rant.

For software engineering, my sense is that the phrase “premature optimization is the root of all evil” has massively backfired. Its from a book on data structures and mainly tried to dissuade people from prematurely write things in assembler. But the point was to free you up to think harder about the data structures to use, not leave things comically inefficient. This context is always skipped when it’s uttered.

Not all fast software is world-class, but all world-class software is fast. Performance is the killer feature.

If you are in engineering, here is a fantastic anecdote. I refer to this account often. It’s a bit subtile, but the implications are massive-

It’s an account of how SQLite became 50% faster, not by doing one specific thing but hundreds of small ones.

SQLite is everywhere today because of this work.

sqlite-users.sqlite.narkive.com/CVRvSKBs/50-fa…

We need the engineers in all companies fight for this more. Product leads are not the right owners of the end performance of the software. This needs to be encoded in the professional pride of the software engineering discipline. Leaders in companies need to encourage it and hold engineering accountable. It’s simply not ok to fritter away the performance of the products for random reasons.

Every user of your products cares exactly as much about latency as engineers do when typing in their terminal. They just don’t have the words to describe what they don’t like about the experience and neither should they.

Jerry Seinfeld on why he still works hard: "Because the only thing in life that's really worth having is good skill. Good skill is the greatest possession. The things that money buys are fine. They're good, I like them, but having a skill - I learned this from reading Esquire magazine. They did one issue about Mastery. This is a very Zen Buddhist concept: pursue mastery. That will fulfill your life. You will feel good. I know a lot of rich people. They don't feel good - as you think they should and would - they don't. They're miserable. So I work because if, you don't, in standup comedy, if you don't do it a lot, you stink."