This little trove of photos is entirely Anil Dash’s fault. He recently made a blog post titled How Markdown Took Over the World with a “hero image” featuring a 2002-era iMac setup — the G4 model, which Jony Ive described as the answer to the question “What computer would the Jetsons have had?”
I stumbled across Anil’s article on a Sunday afternoon while doing some home office housekeeping, and thought, Wow, there’s a blast from the past, which was quickly followed by Hey, didn’t I get someone to take a photo of me with one of those iMacs when they were only days old?
Unlike some people, I’ve been archiving my digital photos since I bought my first digital camera in 1998, so it didn’t take long for me to dig up that iMac G4 photo:
I’d normally tell you to click on this photo to view it at full size, but this is full size — a glorious 640 by 480 pixels!
This photo is from a set that I shot while attending the O’Reilly 2002 ETCon, which took place at the Westin Hotel in Santa Clara. It was a little different from the previous year’s edition of the conference, what with everything being scaled down; even the name had been shortened from the original “Emerging Technologies Conference” down to “ETCon.”
(One of us — I don’t recall whom — engaged in a little gallows humor and quipped “Maybe we should call this the Receding Technologies Conference.”)
This was early 2002, when the dot-com bubble burst had grown into a full meltdown. A lot of us had lost our jobs when the companies we worked for had imploded, but most of us had saved enough to attend a couple of conferences, partly to look for our next gigs, and partly because we couldn’t not go and not see our friends and peers.
One of the nice things about this pared-down conference was that it felt a little more personal. There were more opportunities to just hang out as friends and enjoy some “down time” — or what passes for down time when you’re young and have programming skills, spare time, and lots of ideas — with each other.
The geekist lobby on Earth
We did a lot of hanging out in the hotel lobby, which was pretty much a gathering of a lot of Web 2.0 names you might remember from way-back-when, and some of whom are still working away these days.
This was an awesome “hallway track.”
Cory Doctorow.Damian Stolarz and Wes Felter.Matt Jones, Quinn Norton, and more!
Cory Doctorow.
Jim McCoy.Jim McCoy.
The “Virtual Concierge” at the hotel. There was a concierge desk, but instead of a person behind the desk, there was a monitor and camera, through which you spoke to a concierge based in the Philippines. Hello, kababayan!
Out for dinner and to see the new iMacs
A whole bunch of us decided to go into town for an unofficial dinner one evening, which included a run to the Apple Store to see the new iMac G4s. I took a cab with Zooko, Wes Felter, and Aaron Swartz.
Zooko.Wes Felter and Aaron Swartz.Lisa Rein tries on my cowboy hat while Jillzilla looks on.Quinn Norton, Damian Stolarz, and Jillzilla.Justin Chapweske.Quinn Norton, Matt Jones, ???Wes Felter at the Apple Store.Joey de Villa at the Apple Store with the G4 iMacs.
I suppose I actually attended some conference sessions…
…the only evidence I have of that are these photos.
Clay Shirky, ???, Cory Doctorow.
Nelson Minar and ??? (trying to remember his name).
Attack of the House Party / Attack of the Clones
???, Danny O’Brien, Meg Hourihan.
Star Wars: Attack of the Clones had just been released around the time of the conference, and a number of us decided to go catch the movie. It was decided that we’d pre-game at Quinn Norton’s and Danny O’Brien’s house. This worked well for Aaron, as it wasn’t an age-restricted event at a bar.
In the foreground: Aaron Swartz chats with Lucas Gonze. In the background: Kevin Burton looks on and Matt Jones peeks suspiciously through a doorway.Meg Hourihan, Bram Cohen, Quinn Norton, Cory Doctorow and Jason Kottke.Bram Cohen, Danny O’Brien, Kevin Burton.
After hanging out at the house for a little bit, we made our way to the theatre. We somehow managed to get tickets despite the crowds and our late arrival.
We’d broken up into smaller groups, and Aaron was with me. There were very few seats left, but the front row was still free.
“Front row, then?” I asked Aaron, and he said “Sure.” We took a couple of seats on the left side.
There was still a fair bit of time to kill before the coming attractions came on, never mind the film.
“Dare you to play something,” Aaron said, pointing at my accordion.
We went to see the just-released Attack of the Clones that night!
When the film started, I wanted to get a picture for my blog review. As I pulled out my camera, I said “Keep an eye out for ushers” as I snapped a picture of the opening crawl.
We both got a great laugh out of an all-caps line in the crawl, “CLONE ARMY OF THE REPUBLIC,” and for the next few months, it became a catchphrase for us in IRC chats: “PEER-TO-PEER ARMY OF THE REPUBLIC”, “BOY BAND ARMY OF THE REPUBLIC”, “UNDERPANTS ARMY OF THE REPUBLIC”, and so on.
Final day
On the final day of ETCon, we ditched the lobby and went poolside…
Quinn Norton.Nelson Minar, Clay Shirky, Wes Felter.Cory Doctorow wearing his goggles. Alas, no cape or hot-air balloon.Damien Stolarz in Cory’s goggles.Sam Ruby, Nelson Minar, Cory Doctorow, Wes Felter, Joey de Villa.
Recently, in a fit of annoyance about some big players in tech and particularly AI, I posted the slide featured above across my usual social media channels, and it caught the attention of none other than Jeff Atwood, co-founder of Stack Exchange and Stack Overflow, creator of the Discourse discussion forums platform, and of course, author of the blog Coding Horror.Nowadays, he’s doing a lot of good deeds.
He asked for a link to the slide deck, which I shared, after which he said that I should post in a more web-friendly format. I thought it was a good idea, and I told him I’d do just that.
Since mine was going to be the third of three presentations at an event that was also a cocktail party, I decided to go a little less tech-heavy, a little more “here’s what you might need to brush up on,” and pay a lot of attention to a topic that’s usually a footnote in AI presentations: ethics.
Here are the slides and notes. My presentation was the third of three given at the SocialCode x Tampa event held on September 7, 2023. There’s not much online about it — most of what you can find was written by Yours Truly, including this “save the date” blog post.
Fun fact: It was at SocialCode x Tampa that I hastily improvised what is now the “official unofficial” AI anthem:
The slides and notes
Hello, and welcome to the slide deck from my presentation at the SocialCode x Tampa event that took place on Thursday, September 7, 2023 at Hyde Park Public Studio in Tampa, Florida, USA!
The mandatory speaker bio slide. If you were wondering, my headshot comes from Vadim Davydov, an amazing headshot photographer based here in Tampa Bay. You can find out more about him at https://vadimdavydov.com/.
[ Editorial note: At the time of writing, I am NOT a Senior Developer Advocate for Auth0 by Okta, but I could be YOUR Senior Developer Advocate, because I’m looking for work!If you’re curious, here’s my LinkedIn profile, and here’s my developer relations portfolio. ]
What’s the point of AI? Why do we want to build machines that could render us useless or might even try to eliminate us in the first place?
I like Steve Jobs’ metaphor for computers: “Bicycle for the mind.”
Jobs liked to reference a Scientific American article featuring a study of the energy-efficiency of motion for various species. Unaided, humans were pretty inefficient — until you put a human on a bicycle. When that happens, the human becomes the most energy-efficient species.
Just as a bike augments a human’s body, a computer augments a human’s mind.
From greek mythology, a centaur is like a horse, but with a human head, arms, and torso where the horse’s head would normally be.
The reverse version wasn’t as popular.
“Centaur chess” is Kasparov’s term for a human playing chess with the assistance of a chess program. The human still makes the final decision about chess moves, and uses the computer to try out possible moves and see what the consequences could be.
The minotaur is the closest thing to reverse centaur.
In AI terms, the minotaur is the “opposite” of the centaur. The centaur is a human head in charge of a non-human body, while the minotaur is a non-human head directing a human body. For a “minotaur” AI/human combination, the AI directs the human.
I’d much rather be in a “centaur” relationship with an AI than in a “minotaur” one. I feel bad for soldiers of the future, because there are military thinkers suggesting that the future of warfighting isn’t soldiers ordering machines around, but machines ordering soldiers around.
We’ve had “minotaur warfighting” scenarios in science fiction for decades. One of the best known examples is the Star Trek (original series) episode where Starfleet equips the Enterprise with an AI that then proceeds to attack all other ships, even though those ships are on the same side. Captain Kirk had to talk the AI into committing suicide!
Ideally, I’d like my relationship with AI to be centaur-like when possible, and only minotaur-like when necessary.
In fact, I’d like to be the best Centaur: Chiron. He was respected and taught many Greek gods!
Note the look on Achilles’ face in the painting. It looks like he’s saying “I don’t think that this is an appropriate student-teacher relationship.”
And here we get to the main point of this presentation — how does one get in on some of this sweet, sweet AI action and some of that sweet, sweet AI money? By being in a centaur relationship with AI, or better still, by building AI. How?
I want to flip the standard AI presentation upside-down and start with what’s often a footnote: Ethics. You’ll find that there are some AI “thought leaders” actively fighting against AI ethics, claiming that it’s an unnecessary restriction of progress (and I’ll name and shame one in a few slides).
Let me cite the infamous Boston Housing Dataset. It’s been around for a long time and has been one of the standard datasets that beginning data and AI scientists use; it’s featured in all sorts of tutorials on working with real-world data.
The dataset was used for a 1970s paper titled Hedonic housing prices and the demand for clean air. “Hedonic” is a fancy-schmancy word for “having to do with pleasure,” and the paper was basically about how air pollution affected housing prices.
Air quality in the US used to be terrible in the 1960s and 1970s. It’s so much better today thanks to the EPA and the Clean Air Act, which in turn are thanks to…President Richard Nixon!
If you’re going to do data science or AI development in Python, you’ll eventually use scikit-learn, a library with all sorts of functions for machine learning.
To help people who are new to data science or scikit-learn, it comes with a number of built-in datasets for you to work on, and one of them was Boston Housing Prices.
Here’s a table showing the first 5 rows of the dataset. It seems innocent enough at first glance…
Like many datasets, the columns names are short. The CRIM column is short for “Per capita CRIMe” rate by town.
A little suspicious. I’m going to mark it as such. (Remember, this is about the relationship between housing prices and air quality.)
Then there’s the column named LSTAT, which is short for “Lower STATus,” where “lower status” means “a combination of people who didn’t get at least some time in high school and blue collar guys.”
Again: Housing prices. Air quality.
That’s not just suspicious, but HARVARD suspicious. The worst kind of suspicious.
And then there’s the column simply known as “B”, which is described above. Read the description — then READ IT AGAIN, just to be sure:
“Black proportion of population. At low to moderate levels of B, an increase in B should have a negative influence on housing value if Blacks are regarded as undesirable neighbors by Whites. However, market discrimination means that housing values are higher at very high levels of B. One expects, therefore, that a parabolic relationship between proportion Black in a neighborhood and housing values.”
To quote M Carlisle, who summarized the big problem with the Boston Housing Prices dataset in the article racist data destruction?
This actually is a parameter in the model to modulate house pricing for systemic racism. When I say ‘systemic racism’ here, I mean this mathematically. This is a term, in a statistical model to predict housing prices, that accounts for racism as a factor in pricing. If this is used to predict, or even influence future models by its very existence, then systemic racism will continue to be a pricing factor.
This bit earns more than a mere SUS. I’m giving it the Ice Cube WTF.
Now if you look at the values in the “B” column, they’re in the 390s. Does this mean that B is reporting the Black proportion of the population expressed as units of thousandths, where 0 = a complete lack of Black people and 1000 = everyone is Black?
Sadly, no. Instead, B contains a calculated value — it’s the result of the equation y = (x – 0.63)^2 (In many programming languages, “^” is the exponent operator. For example 3^2 means “3 raised to the power of 2,” which is 9.)
y is the resulting value that goes into the B column. x is the actual proportion of Black people in the area, as you can see that x has to be between 0 and 1 (that’s what the “0 < x < 1” part means).
Equations of the form y = x^2 describe upward-opening parabolas. With a basic y = x^2 parabola, the lowest point of the parabola happens when x is 0. With the formula for the parabola above — y = (x – 0.63)^2 — the parabola gets shifted to the right, so that the lowest point is when x = 0.63. In this case, the function says that house prices:
Are highest when a neighborhood is all-White
Lowest when a neighborhood is 63% Black
Climb as a neighborhood becomes more than 63% Black, but only up to a point that’s less than half the house prices in an all-white neighborhood. The graph hits this limit because x has to be between 0 and 1 (you can’t have fewer Black people than 0, or more Black people than 100%).
Another problem with this equation is that it loses original data. For almost every value of y (except the parabola’s lowest point), there are two possible values for x: the one on the left side of the parabola, and another one on the right side.
Once again, I remind you: this was for a research paper about AIR QUALITY affecting HOUSE PRICES.
There are other issues with this data, and they’re covered in the essay racist data destruction?.
This was also the 1970s. We’re talking barely a decade after the Civil Rights Act of 1964, the deaths of both Martin Luther King and Malcolm X, and the 1967 court case of Loving v. Virginia, whose ruling finally made interracial marriage legal in all 50 states. We still have a way to go, but wow, was it downright atrocious then.
But we’ve stopped doing unethical science since the bad old days, right? RIGHT?
Ummm…no.
Alas, one of the big names in AI, Marvin Minsky — a cognitive scientist and computer scientist at MIT whose artificial intelligence work is foundational — held a couple of conferences on Jeffrey Epstein’s private island. And yes, this was AFTER Epstein was registered as a sex offender after pleading guilty in 2008 to soliciting a minor for prostitution. By that point, if you were going to the island, you at least knew what “recreational activities” he offered.
Virginia Giuffre, whom Epstein had trafficked for Prince Andrew, testified that Epstein’s associate Ghislaine Maxwell that Maxwell “directed” her to have sex with various people, including Minsky. At the time of the alleged incident, Giuffre was 17 and Minsky was…73.
[ Editorial note: I gave this presentation on September 7, 2023, when Giuffre was still alive. She died by suicide a year and a half later, on April 25, 2025. ]
George Church, notable biologist at Harvard — what is it with these Harvard and MIT nerds? — took meetings and calls with Epstein, even after the prostitution-with-a-minor deal became public knowledge. Unlike Minsky, he fully and publicly apologized: “There should have been more conversations about, should we be doing this, should we be helping this guy? There was just a lot of nerd tunnel vision.” Remember that phrase, “nerd tunnel vision,” and learn not to get trapped in it. There is more to our area of study than just the study itself, but how what we do affect others.
Speaking of nerd tunnel vision…
Ugh, this guy. First famous for helping create Mosaic, the OG graphical browser, then forming the VC firm Andreesen Horowitz, and now super-big AI proponent, but against anything that might slow down its development, even things like ethics.
That damned Silicon Valley mindset. I swear, there should be some kind of system that requires you to get punched in the groin every time you open a copy of anything written by Ayn Rand.
As tech and AI get increasingly intertwined with everyday life, we need to keep this in mind.
Check this stuff out, especially Timnit Gebru’s writing. She was the AI scientist fired by Google for refusing to retract a paper that asks whether enough thought has been put into the potential risks associated with developing large language model-based AIs and strategies to mitigate these risks. That paper is On the Dangers of Stochastic Parrots.
Yes, you can write all sorts of apps that make calls to some AI that provides an API, and there are a lot of use cases where that’s more than enough.
But what if you want to build AIs yourself? If that’s the case, you’ll need to brush up on some math. There are a number of AI libraries that do a lot of the math for you, but you should at least learn the underlying principles. It’s like using a calculator or spreadsheet even though you know arithmetic.
The text on this slide captures the general idea.
If you’ve written code to draw things onscreen, you’ve probably worked with x-, y-, and possible z-coordinates, and that’s part of linear algebra, which is the algebra of lines and planes.
Here’s a simple example of linear algebra being used to perform classification, which is just a fancy way of saying “telling things apart.”
Suppose you’re working with two different kinds of garden creatures: caterpillars and ladybirds. Caterpillars are long creatures, and not very wide. On the other hand, ladybirds are wide creatures, but not very long.
If you took the length and width measurements of caterpillars and ladybirds and plotted them on 2-D graph, with the x-axis representing the creature’s width and he y-axis representing the creature’s length, you’d get a graph like the one above. In this graph, the caterpillar measurements are clustered around the upper left, representing values where the creature is long, but not wide. The ladybirds are clustered around the lower right, representing values where the creature is wide, but not long.
We can use linear algebra to find the equation of a line that divides the measurements into two different zones: one for caterpillars, and one for ladybirds. If a creature falls on the “caterpillar” side of the line, we classify it as a caterpillar. If it falls on the “ladybird” side, we classify it as a ladybird.
The text on this slide captures the general idea.
Speed is the rate of change of distance. You can see it in the way we express speed — in terms of distance over time (for example, 55 miles per hour).
The text on this slide captures the general idea.
The text on this slide captures the general idea.
It’s not every day I can Rickroll an entire room!
The text on this slide captures the general idea.
These books are friendly introductions to the kind of math you’ll want to know if you want to developer AI applications:
If you need an entertaining introduction to AI, what it can do, and more importantly, what it does terribly, hilariously, wrong, you’ll want to check out You Look Like a Thing and I Love You. The book gets its title from an AI-generated pickup line.
It’s written by Janelle Shane, author of the blog AI Weirdness, which is also a worthwhile read!
One of the big challenges of AI is explainability, which is the ability to show how an AI came to a decision in a way we can understand and trust. The challenge is that simple AI systems are explainable, but aren’t terribly impressive. The more complex an AI becomes, the less explainable it is.
The first AIs were rules-based systems and were essentially giant collections of “IF (this happens) THEN (respond this way)”-style commands.If you could write rules for every possible situation, this style of AI would work well, The problem is, you can’t. But these systems are quite explainable.
Of course, deep learning systems aren’t rules-based — or at least we’re not providing the rules. Instead, they’re based on artificial neural networks (we often shorten the term to plain old “neural networks”), which are based on the biological neural networks found in animals, including humans. They’re networks of nerve cells, a.k.a. neurons, the building blocks of the brain and nervous system. Individually, they don’t do much, but when arranged into network, they can perform complex calculations.
Here’s a closer-up look look at a biological neuron…
…and here’s a technological neuron. The dendrites are inputs that take in numbers, each of which is then multiplied by a weight value. These results are totalled into a sum, which is then run through an activation function, which activates only if that sum crosses a specific threshold. If the sum does cross that threshold, the neuron produces an output value, which is the equivalent of a biological synapse firing.
It’s things like these that make me think “I’m not sure I like being in this episode of Black Mirror.”
However, these two books do cover some interesting topics and seem worth checking out. Maybe buy them under a different identity!
And yes, I’ll talk a little bit about programming. Not too much, because my time’s nearly up, and because there are more than enough OTHER AI presentations on writing code.
The text on this slide captures the general idea.
Objective-C, while a pretty good programming language (and in my opinion, preferable to C++), languished in obscurity through the 1980s and 1990s. The only platform that really used it was NeXTSTEP, the operating system for NeXT Computer, which Steve Jobs founded after getting kicked out of Apple.
Objective-C might have stayed obscure if not for Apple purchasing NeXT in 1996 and Steve Jobs taking over Apple after CEO Gil Amelio was ousted in 1997. This led to the old Apple “System” OS being replaced with OS X, which was NeXTSTEP with an Apple coat of paint on top of it. OS X was the basis for iOS, and iOS drove adoption of Objective-C to never-before-seen levels.
It was the TIOBE Programming Language Index’s language of the year for 2011 and 2012 and peaked in popularity in 2014 — the year Swift (the current programming language for Apple systems) was introduced. Even today, there are still vestiges of NeXTSTEP in macOS, iOS, and other Apple operating systems — you’ll still see classes whose names begin with the letters “NS,” which is short of NeXTSTEP.
Python was introduced in 1991 but was constantly overshadowed by other programming languages — first by Perl in the late 1990s, when web developers adopted it heavily before switching to PHP. Then, in the late 2000s, it was eclipsed by Ruby, thanks to Ruby’s killer app, Ruby on Rails.
However, the web revolution brought about by Perl, PHP, and Python drove the data science revolution of the 2010s — and Python was thriving in data science, thanks to its easy-to-learn syntax and extensive math and scientific computing libraries. That same data science revolution also meant that for the first time, it was easy to access lots of data to train neural networks, and AI grew in leaps and bounds. Today, Python is the top language on the TIOBE Index, and was TIOBE’s programming language of the year in 2021.
I have a “programming language time-investment strategy” that goes like this:
I devote 70% of my time to the programming languages that pay the bills. In my current role as a developer advocate at Okta specializing in mobile development, these languages are Swift and Kotlin.
I devote 20% of my time to the programming languages that look like likely next things. For a while, that was, and continues to be Python, although now I’m more focused on Python for AI.
I devote 10% of my time to the programming languages that just look like fun, and who knows, it may pay off! Once upon a time, Python was in this slot. These days, those languages are…
Julia — it’s supposed to have:
C’s speed
Python’s flexibility
Ruby’s dynamism
R’s statistics capability
Matlab’s linear algebra capability
Mojo — a very new language created by the same person who created Swift (the Apple programming language), which has these benefits:
It’s a superset of Python — that is, valid Python code is also valid Mojo code. You can stick to writing only Python code in Mojo, but then you’d be missing out!
Because it’s a superset of Python, it can use Python’s libraries.
Mojo is superfast. It’s designed to take advantage of the hardware it’s running on, including GPUs. It also has built-in parallelization, autotuning, types and type-checking, and all sorts of optimizations to make it faster.
Both are great for AI. I’m paying a little more attention to Mojo at the moment, but at the speed at which things are moving these days, you have to be ready to pivot!
At last, the conclusion!
When ChatGPT was released in late November 2022, I showed it to friends and family, telling them that its underlying “engine” had been around for a couple of years. The GPT-3 model was released in 2020, but it went unnoticed by the world at large until OpenAI gave it a nice, user-friendly web interface.
That’s what got me thinking about my thesis that 2020 might be the start of a new era of initially-unnoticed innovation. I started looking backwards to the previous industry-changing leap: the iPhone in 2007, which created the age of the smartphone, which gave us ubiquitous computing and the mobile internet.
13 years before 2007 is 1994, the year the Netscape Navigator browser was introduced, and created the age of the web.
13 years before 1994 is 1981, the debut of the IBM PC, the machine that put desktop computers in more offices and homes than any other.
13 years before 1981 is 1968, when The Mother of All Demos took place — Douglas Englebart’s demonstration of what you could do with computers, if they got powerful enough. He demonstrated the GUI, mouse, chording keyboard, word processing, hypertext, collaborative document editing, and revision control — and he did it Zoom-style, using a remote video setup!
It’s 1968, 1981, 1993, and 2007 all over again!
How do you capitalize on this opportunity as a centaur?
STOP DOING AVERAGE SHIT, like my mom says — but minus the swear word.
Average is staying at home and watching the new live-action “One Piece” on Netflix (which is getting record-breaking viewer numbers, by the way). Your being here to learn about AI and meet with fellow AI aficionados is NOT average. Keep it up — coming to events like this is how you learn, share ideas, and stay inspired!
Be sure to check out these local AI and AI-related meetups:
There are some great conferences here and nearby. Even people from outside the US are choosing Tampa Bay for their conferences, like Civo Navigate back in February. If you missed that, you missed a great one.
And for the next 2 years, PyCon US — the world’s biggest Python conference — is happening in Pittsburgh.
And last but not least — the best way to be a centaur is to just build things! Just follow the clean version of Florida’s unofficial state motto: “Fool around and find out!”
On Saturday, I had the honor of speaking on the Careers in Tech panel at TechX Florida, which was organized by USF’s student branch of the IEEE Computer Society.
…and I enjoyed performing the “official unofficial song of artificial intelligence” at the end of the panel:
Reasons to be optimistic 2025
During the panel, a professor in the audience asked an important question on behalf of the students there: In the current tech industry environment, what are the prospects for young technologists about to enter the market?
I was prepared for this kind of question and answered that technological golden ages often come at the same time as global crises. I cited the examples from this book…
…Thank You for Being Late, by Thomas Friedman, who proposed that 2007 was “one of the single greatest technological inflection points since Gutenberg…and we all completely missed it.”
The reason many people didn’t notice the technological inflection point is because it was eclipsed by the 2008 financial crises.
During the dark early days of the COVID-19 pandemic and shutdown, the people from Techstars asked me if I could write something uplifting for the startupdigest newsletter. I wrote an article called Reasons for startups to be optimistic, where I cited Friedman’s theory and put together a table of big tech breakthroughs that happened between 2006 and 2008.
In answering the professor’s question, I went through the list, reciting each breakthrough. The professor smiled and replied “that’s a long list.”
If you need a ray of hope, I’ve reproduced the list of interesting and impactful tech things that came about between 2006 and 2008 below. Check it out, and keep in mind that we’re currently in a similar time of tech breakthroughs that are being eclipsed by crises around the world.
Originally started in 2003 and bought by Google in 2005, Android was at first a mobile operating system in the same spirit as Symbian or more importantly, Windows Mobile — Google was worried about competition from Microsoft. The original spec was for a more BlackBerry-like device with a keyboard, and did not account for a touchscreen. This all changed after the iPhone keynote.
In case you don’t remember, Steve Jobs’ original plan was to not allow third-party developers to create native apps for the iPhone. Developers were directed to create web apps. The backlash prompted Apple to allow developers to create apps, and in March 2008, the first iPhone SDK was released.
Azure
Azure, Microsoft’s foray into cloud computing, and the thing that would eventually bring about its turnaround after Steve Ballmer’s departure, was introduced at their PDC conference in 2008 — which I attended on the second week of my job there.
Bitcoin
The person (or persons) going by the name “Satoshi Nakamoto” started working on the Bitcoin project in 2007.
The end of the year 2007 marks the first time that the cost of genome sequencing dropped dramatically — from the order of tens of millions to single-digit millions. Today, that cost is about $1,000.
In September 2006, Facebook expanded beyond universities and became available to anyone over 13 with an email address, making it available to the general public and forever altering its course, along with the course of history.
Energy technologies: Fracking and solar
Growth in these two industries helped turn the US into a serious net energy provider, which would help drive the tech boom of the 2010s.
GitHub
Originally founded as Logical Awesome in February 2008, GitHub’s website launched that April. It would grow to become an indispensable software development tool, and a key part of many developer resumes (mine included). It would first displace SourceForge, which used to be the place to go for open source code, and eventually become part of Microsoft’s apparent change of heart about open source when they purchased the company in 2018.
Hadoop
In 2006, developer Doug Cutting of Apache’s Nutch project, took used GFS (Google File System, written up by Google in 2003) and the MapReduce algorithm (written up by Google in 2004) and combined it with the dataset tech from Nutch to create the Hadoop project. He gave his project the name that his son gave to his yellow toy elephant, hence the logo.
By enabling applications and data to be run and stored on clusters of commodity hardware, Hadoop played a key role in creating today’s cloud computing world.
Intel introduces non-silicon materials into its chips
January 2007: Intel’s PR department called it “the biggest change to computer chips in 40 years,” and they may have had a point. The new materials that they introduced into the chip-making process allowed for smaller, faster circuits, which in turn led to smaller and faster chips, which are needed for mobile and IoT technologies.
On January 9, 2007, Steve Jobs said the following at this keynote: “Today, we’re introducing three revolutionary new products…an iPod, a phone, and an internet communicator…Are you getting it? These are not three separate devices. This is one device!”
The iPhone has changed everyone’s lives, including mine. Thanks to this device, I landed my (current until recently) job, and right now, I’m working on revising this book.
iTunes sells its billionth song
On February 22, 2006, Alex Ostrovsky from West Bloomfield, Michigan purchased ColdPlay’s Speed of Sound on iTunes, and it turned out to be the billionth song purchased on that platform. This milestone proves to the music industry that it was possible to actually sell music online, forever changing an industry that had been thrashing since the Napster era.
Kindle
Before tablets or large smartphone came Amazon’s Kindle e-reader, which came out on November 19, 2007. It was dubbed “the iPod of reading” at the time.
You might not remember this, but the first version didn’t have a touch-sensitive screen. Instead, it had a full-size keyboard below its screen, in a manner similar to phones of that era.
Macs switch to Intel
The first Intel-based Macs were announced on January 10, 2006: The 15″ MacBook Pro and iMac Core Duo. Both were based on the Intel Core Duo.
Motorola’s consistent failure to produce chips with the kind of performance that Apple needed on schedule caused Apple to enact their secret “Plan B”: switch to Intel-based chips. At the 2005 WWDC, Steve Jobs revealed that every version of Mac OS X had been secretly developed and compiled for both Motorola and Intel processors — just in case.
In 2007, Netflix — then a company that mailed rental DVDs to you — started its streaming service. This would eventually give rise to binge-watching as well as one of my favorite technological innovations: Netflix and chill (and yes, there is a Wikipedia entry for it!), as well as Tiger King, which is keeping us entertained as we stay home.
Python 3
The release of Python 3 — a.k.a. Python 3000 — in December 2008 was the beginning of the Second Beginning! While Python had been eclipsed by Ruby in the 2000s thanks to Rails and the rise of MVC web frameworks and the supermodel developer, it made its comeback in the 2010s as the language of choice for data science and machine learning thanks to a plethora of libraries (NumPy, SciPy, Pandas) and support applications (including Jupyter Notebooks).
I will always have an affection for Python. I cut my web development teeth in 1999 helping build Givex.com’s site in Python and PostgreSQL. I learned Python by reading O’Reilly’s Learning Python while at Burning Man 1999.
Shopify
In 2004, frustrated with existing ecommerce platforms, programmer Tobias Lütke built his own platform to sell snowboards online. He and his partners realize that they should be selling ecommerce services instead, and in June 2006, launch Shopify.
Spotify
The streaming service was founded in April 2006, launched in October 2008, and along with Apple and Amazon, changed the music industry.
Surface (as in Microsoft’s big-ass table computer)
Announced on May 29, 2007, the original Surface was a large coffee table-sized multitouch-sensitive computer aimed at commercial customers who wanted to provide next generation kiosk computer entertainment, information, or services to the public.
2007 was the year that networking switches jumped in speed and capacity dramatically, helping to pave the way for the modern internet.
Twitter
In 2006, Twittr (it had no e then, which was the style at the time, thanks to Flickr) was formed. From then, it had a wild ride, including South by Southwest 2007, when its attendees — influential techies — used it as a means of catching up and finding each other at the conference. @replies appeared in May 2007, followers were added that July, hashtag support in September, and trending topics came a year later.
After performing poorly financially, the husband and wife cofounders of VMWare — Diane Greene, president and CEO, and Mendel Rosenbaum, Chief Scientist — left. Greene was fired by the board in July, and Rosenbaum resigned two months later. VMWare would go on to experience record growth, and its Hypervisors would become a key part of making cloud computing what it is today.
Watson
IBM’s Watson underwent initial testing in 2006, when Watson was given 500 clues from prior Jeopardy! programs. Wikipedia will explain the rest:
While the best real-life competitors buzzed in half the time and responded correctly to as many as 95% of clues, Watson’s first pass could get only about 15% correct. During 2007, the IBM team was given three to five years and a staff of 15 people to solve the problems. By 2008, the developers had advanced Watson such that it could compete with Jeopardy! champions.
You probably know this device better as the “One Laptop Per Child” computer — the laptop that was going to change the world, but didn’t quite do that. Still, its form factor lives on in today’s Chromebooks, which are powered by Chrome (which also debuted during this time), and the concept of open source hardware continues today in the form of Arduino and Raspberry Pi.
YouTube
YouTube was purchased by Google in October 2006. In 2007, it exploded in popularity, consuming as much bandwidth as the entire internet did 7 years before. In the summer and fall of 2007, CNN and YouTube produced televised presidential debates, where Democratic and Republican US presidential hopefuls answered YouTube viewer questions.
This Saturday, November 8, I’ll be at the TechX Florida 2025 AI Conference at USF, on the Careers in Tech panel, where we’ll be talking about career paths, hiring expectations, and practical advice for early-career developers and engineers.
This conference, which is FREE to attend, will feature:
AI talks from major players in the industry, including Atlassian, Intel, Jabil, Microsoft, and Verizon
Opportunities to meet and network with companies, startups, and techies from the Tampa Bay area
The Careers in Tech panel, featuring Yours Truly and other experienced industry pros
On Tuesday, two popular tech events take place in Tampa, and you may be wondering which one you should attend. I’ll answer your question by quoting the little girl from that classic Old El Paso commerical:
This week, from Monday through Wednesday, the CyberBay 2025 conference is taking place. Organized by the University of South Florida,Cyber Florida, Bellini Capital, the USF Bellini College of AI, Cybersecurity, and Computing, and the USF Institute for AI+X, CyberBay is where talent, technology, and national security converge to build the future of digital defense.
On the weekend on November 7 – 9, Techstars Startup Weekend comes to Tampa. It’s a hackathon where you’ll compete to build the best startup in a mere 54 hours. There’ll be mentors from industry to help out, and the event is calling for developers, designers, and domain experts.
BSides (a.k.a. “Security BSides”) is a series of cybersecurity conferences that take place all over the world. Here in Tampa Bay, we are especially lucky that we get not one, but two BSides conferences every year, with BSides Tampa happening in the spring and BSides St. Pete taking place in the fall.
BSides gets it name from “b-side,” the alternate side of a vinyl or cassette single, where the a-side has the primary content and the b-side is the bonus or additional content. In 2009, when the Black Hat conference in Las Vegas received way more presentation submissions than they could take on, the rejected presenters (who still had very could presentations; there just wasn’t enough capacity for them) banded together and made their own “b-side” conference that ran in parallel with Black Hat. From that event came BSides.
When is BSides St. Pete happening?
On Friday, October 3, starting at 9:00 a.m., the BSides St. Pete trainings will take place, featuring these topics:
Exposing Coordinated Attacks
TCP/IP Black Ops Training
Introduction to IoT Security: Hands-On CTF Style
RE:Boot – An Introduction to Reverse Engineering Training
What’s my job again?
When Should a Business Use SIEM & SOAR
On Saturday, October 4, starting at 9:00 a.m., the BSides St. Pete conference will take place, featuring all the topics featured on the agenda.
Why attend BSides?
Me at BSides St. Pete 2023. Click to read my report on that event!
BSides conference are great — not only for the information security knowledge you’ll pick up at any given session, but also for the community. Tampa Bay is a cybersecurity hub, and BSides conferences are organized by the community, for the community. You’ll get to meet area folks in the tech and cybersecurity fields, and you might even make friends!
(And yes, I’m talking actual friends here. In fact, after I write this, I’m going to a potluck dinner with friends I met through BSides.)
This little trove of photos is entirely Anil Dash’s fault. He recently made a blog post titled How Markdown Took Over the World with a “hero image” featuring a 2002-era iMac setup — the G4 model, which Jony Ive described as the answer to the question “What computer would the Jetsons have had?”
