Joined April 2011
- Tweets 381
- Following 592
- Followers 65
- Likes 105
1 Photos and videos
Steve Jay retweeted
Jun 12
We need to have a serious conversation about Canada's culture of aggressively resentful mediocrity, which is encouraged by the Laurentians as it helps prop up the rentier feudalism of Canada's moribund political economy.
Success is punished. Ambition is castigated. Achievement goes unrewarded. No one is supposed to get above themselves, no one is allowed to reach for greatness.
In the Canadian system, you are expected to obtain permission for absolutely anything you want to do. Any success that isn't "success" bequeathed from the generous hand of Ottawa's commissars is immediately suspect, and will invariably be sabotaged. That pattern has persisted uninterrupted since Diefenbaker broke the aerospace industry. It's small-minded, small-town provincialism as an ethos of governance. Exhausted late-Soviet Brezhnevism laying like a choking fog over the land.
This pathetic attitude is among the worst things about our political culture in Canada and I cannot reject it enough.
It gets cloaked in the language of progressivism but it is deeply cynical, ugly, and regressive.
Story time!
Growing up, I was hugely inspired by RIM (BlackBerry). It was one of the reasons why I wanted to go to Waterloo.
I thought it was so cool one of the most innovative companies on earth was an hour away from home.
In fact, my program, Nanotechnology Engineering, was able to exist in part due to the philanthropy of Mike Lazaridis, who funded the Institute of Quantum Computing and Nanotechnology (along with the Perimeter Institute for theoretical physics, which is a brilliant asset for the province and country). Balsillie, for his part, has spent tens of not hundreds of millions of his personal wealth on advocacy and institutions to make Canada a better place. But he too was castigated in our media.
Through high school, I saw how Canadas media took an axe to RIM founders (Mike and Jim), and basically cheered on the decline of the business against competition from Apple and Google.
It was a complete disgrace.
Well, in 2013 I got my second co-op job there, just as they rolled out BB10 (the QNX operating system). 6 weeks into my co-op, my entire department was laid off (Modems/Semiconductors).
Nearly every one of my colleagues ended up moving to the US. Some of the most capable talent on earth, poached in weeks. It was loss that was absolutely devastating to witness.
I have no doubt people like Bruce cheered on the spectacle, just like he would cheer the downfall of Shopify if it were to ever happen; despite the champion it’s been for the country, the thousands of good jobs it’s created, and all the spin-off businesses that have created huge wealth for Ontario.
Well let me be clear that I will have none of this nonsense.
28
133
976
26,547
Jun 12
Why does Canada hate success?
Jun 12
You don’t have to love Elon Musk to recognize what this headline says about us.
A country that spends more time criticizing wealth creation than encouraging it sends a clear message to builders: your success is tolerated, not celebrated.
Canada should be the best place in the world to build ambitious companies. Headlines like this make us look like we’re not quite ready for that.
51
Jun 12
What happened @globeandmail? You just changed your editorial standards? The article was published with previous headline approved.
No, you’re just showcasing the hypocrisy of the Canadian left. ‘Canada is an inclusive paradise, unless you’re successful and or wealthy and or male and or white, then you can just f*ck off and die’
Jun 12
The previous headline on this article did not meet The Globe’s editorial standard. It has been replaced.
43
Jun 12
In case you didn’t know: Canada hates success.
Unless it’s fake success propped up by the government for votes from special interest groups.
16
Jun 11
As I continue to grow my obsidian based second brain, I am spending more time on creating a harness around it. So far the harness consists of:
- adding time stamps to entries (Claude code does not have a clock)
- something I call ‘vignettes’ which are observations about the process that serve as inputs for future harness improvements
- cross-tool and cross-session integrations
- the usual suspects: Claude.md, memory.md, user.md
- multiple customer ingestion tools (basically like etl / data cleaning scripts for pulling content from different sources)
- Academic citation sweep
I really enjoy this project, it helps me to think and generate new ideas, keep track of how my ideas evolve, how they’re linked to each other and redundancy.
1
54
Jun 10
Ok fine I’ll just tell you why Multiculturalism doesn’t work and will never work:
Culture is about a set of rules. Rooted in symbolic language. Norms and anti-norms. The ‘way to be’ is more than understood, it is deeply known, and most importantly, felt. In-group members FEEL like members. The out group FEELS like outsiders.
Now when you mash up a bunch of people from a bunch of different cultures, you have different languages, different symbols, different understandings of norms and anti-norms, no one knows how to ‘be’ and no one feels belonged.
So what happens? Central authorities start to create a set of rules using legal and political authority backed by threat of enforcement. The more multi-cultural, the more rules-of-law that are required; the more ‘multi cultural social scaffolding’ gets created. No one feels it, no one knows it.. people sort of, for the most part, understand it.
But I’m sure you can see the problem, the more groups of people you get, the more languages, the more cultures, the more and more difficult it becomes to create this legal/political/social scaffolding, eventually the bureaucracy required to hold the thing together becomes all consuming, the society collapses and the ‘edge cases’ are the reason.
Modern liberalism loves multiculturalism because it requires more and more centralized control, more government and law enforcement. More rules. More money, more taxation… you get the point.
It has never worked. It will never work. We are at the tipping point. The outcome is when, not if.
1
85
Jun 10
The model doesn’t have to hallucinate to introduce risk in the decision making or output generating process.
Models have no accountability.
The value of insurance is heavily discounted at the moment of decision.
My solution? Coming soon.
37
Jun 10
The models are forever prisoners in the cave. Humans are forever outside the cave.
All the value in the cave collapses to perfect economics, all the subjective value accumulates outside the cave.
34
Jun 9
Currently building a comprehensive business plan around this idea.
Mar 10
Audit the AI. That's the job of humans. Only the MOST knowledgeable people who deeply understand their field can know when the AI is full of shit. The value of the AI auditor is commensurate to the value of the risks created by overly confident AI hallucinations.
69
May 31
Working on this knowledge-system thing this week, loading another batch of notes into the LLM, and a weird thing hit me.
This thing has read everything. Every book, every paper, every blog post anyone bothered to write. But it hasn't done anything. It hasn't been anywhere. It hasn't paid for being wrong about anything.
It's a stranger version of Plato's cave.
Plato's prisoners at least watched the shadows on the wall. The LLM has only read other people's descriptions of the shadows.
I'm the one outside.
I'm its eyes, its ears, its hands. Not its brain - it's got plenty of that. It's missing a body.
Everything it knows about the world outside its training data has to come through me - what I noticed today, what I think matters, what I remember about being wrong in 2017 and being right in 2023 and why.
Without that, it's brilliant and staring at a blank wall.
We've got the priest and the supplicant the wrong way around with LLMs. The model isn't the priest. It's the prisoner. I'm the one walking around outside, bringing back compressed versions of what I saw.
There's this weird meta thing happening every time I do it: The model can't go where I'm going. I can never fully bring back what I saw. But what I bring back is the only thing in its world that isn't already a description.
I don't think this gap gets smaller. I think it gets weirder.
62
May 28
This thesis is aging well
May 11
The clean picks-and-shovels play if you're agnostic on winners:
$DDOG $NET $SNOW $MDB
They don't care who wins the SaaS war. They just need more software running in the world.
AI-accelerated development means more apps, more traffic, more data, more monitoring.
Every AI-native 5-person company still runs on infrastructure.
---
not financial advice
98
May 28
Hot take: AI doesn’t make anyone smarter. Smart people do crazy, unbelievable things with AI.
Not smart people use ChatGTP instead of Google to decide what microwave to buy or what hotel to stay at or what that weird symptom is.
Therefore, normies will stay normies and a small percentage of really sophisticated people will create and capture all the value with AI.
50
May 27
AI is the likelihood. You are the prior.
The model has read everything Freud wrote. It hasn't read what you wrote yesterday.
The gap is the only durable edge in a world of identical tools.
Here's the practice. (Part 3 of The Operator's Loop)
83
May 26
Most AI builders are running 95% wasp in a beaver-rewarding environment.
Throwaway prompts. Cursor for one-offs. Cycling every new model. Zero compounding context. Every session starts from zero.
Here's the 5-step audit to fix your ratio.
Part 1 of The Operator's Loop ↓
98
May 26
I've been writing a series called The Operator's Loop.
It's a 4-part recipe for how to read the world the way a builder reads it. Each part is a practice you can run today.
Here's the arc 👇
1/6
1
33
May 26
5/6
Part 4: Play the Rests
What you leave out is doing more work than what you put in. Across music, design, writing, strategy - and especially the AI products being built right now.
Here's why the negative space is the work.
1
28
May 26
6/6
Read in order. Part 1 here → x.com/robosauce/status/20589…
New part dropping every few days. The whole series is a loop: read environment → find mispricing → bank judgment → recognize unsaid.
Tag me if any of them lands.
28