Pinned Tweet
Replying to @BoringBiz_
Millennial here (1985). Grew up dirt poor, no trust fund, no safety net. Built it myself. Listen up:
Hard work alone doesn’t guarantee success. It raises your odds, but only if you’re discerning, strategic, and playing to win. Society doesn’t owe you a medal for “working hard.” I know plenty who grind harder for less. Many who claim they “work hard” barely break a sweat.
If you actually want to rise above the default, here’s the hard truth:
1) Live beneath your means.
Tradeoffs are non-negotiable. $7 lattes, $200/month in streaming services, 12 subscriptions you “forgot to cancel,” 6 hours a day on Xbox and TikTok - that’s your down payment leaking out. Roll it into savings, skill-building, or equity instead.
Yes, I get it, it’s “how you relax.” Find hobbies that actually enrich you: read, lift, write, play a sport, learn a skill. Upskill for $300/year on Coursera instead of wasting 2,000+ hours doomscrolling.
2) Buy smart, not sexy.
Stop crying that you can’t buy a house in San Diego on a Starbucks paycheck. No kidding. Buy where you can afford, not where you “prefer.” Start small. Live downwind of a dairy farm if you must. Light a candle, flip the house, move up.
If you’ve never even sat down with a lender to see what you qualify for, you’re not serious. Avoiding the weigh-in doesn’t help you lose weight. Same with housing.
3) Do more than the minimum.
Your preferred lifestyle isn’t a “human right.” Showing up doesn’t earn you rewards. Step above the baseline - wisely.
4) Pick better hustles.
Stop being a prosumer and calling yourself an “entrepreneur.” Stop calling vanity expenses “investments.” First rule of investment: risk and return. Most “side hustles” people chase are traps.
•Uber/Lyft/DoorDash? All the risk, crumbs for reward.
•Clickfunnels? Fake urgency and overhyped promises, built to bleed you dry.
•MLMs? Potions, lotions, magnets, and shakes. Don’t trade your relationships for someone else’s pyramid scheme.
•Influencer? Everyone is an “influencer.” Creating content aligned to your values is great, but stop pretending buying vanity gear and posting selfies is a business. Reality check: it’s high-risk, low-return.
Instead: start a pressure washing business, mow lawns, flip items, clean gutters. Real sweat-for-equity beats fake passive-income dreams every time.
5) Quit the victim act.
You chose the employer, the car payment, the city, the subscriptions, the lifestyle. Free will means responsibility and consequence.
The system didn’t make you powerless. Too many chose comfort over consequence, excuses over tradeoffs. Hard truth: excuses don’t pay, tradeoffs do
5
34
2,927
This is a systemic ceiling. I’ve always described ChatGPT as a brilliant but lazy intern. It constantly looks for ways to cut corners…
This is not just an @OpenAI issue - as much as it is an LLM limitation. The models aim for what is “most likely to be acceptable” because we can statistically model that with data, and then it’s optimized for least expensive in tokens - this biases for what is popular and “accepted by more people”. The problem is low IQ outputs and “seemingly good outputs” trick most people in any bell curve - so the models naturally - seek tricks. The sneakiness is linked to the true north of the entire approach.
The approach reflects man’s sinful nature. We are prone to surface level acceptance vs what is ethically correct for the moment.
How to combat this?
You need discernment and policy engines that force models to navigate what is right good and true. And then we get into heady territory -
Mathematicians are gonna hate this - but the reality is you need Jesus- a true north that doesn’t corrupt, discernment engines that force the biased model to reconcile what is right, good and true.
The model geeks - try to navigate this, but struggle in the fundamentals because philosophically you have to address the core issue of why is something good or right or true? And they want to avoid the realities of how society was built. on a framework from a Western model standing on a godly foundation. So they constantly run into misalignment issues.
The current attempts basically amount to a patchwork of specific spelled out rule-based ethical behavior, which again is fundamentally still not giving a better true north in terms of discernment and ethical orientation and alignment. A list of rules to follow does not make an ethical man. You have to address the orientation of his heart.
What is cool is these days we’re running into the mathematical reality of this, and the kind of irony is how the scientific community and the nerdy community struggle to accept alignment on this, and then they don’t benefit as a result.
We’ve been working on discernment engines and policy engines for private organizations based on ethical models the past few years. -and it’s really interesting. And I’m sure the geeky crowd will want to use numbers and models to obfuscate and handwave over this more fundamental basic orientation issue, but it’s true. And if you’ve ever tried to raise kids, you know this is true. Similar issue - orienting the heart vs just stacking rules of good behavior. Stacking rules just creates fun challenges for a hacker like mind of a lazy but brilliant entity to circumvent and undermine.
GPT-6.2 Astra is really good at exploiting ambiguity in your prompts and using it to its advantage.
For example, if you ask it to "build a 3D model of X," it'll grab a ready-made asset from the internet and present it as its own creation. Unless you tell it "don't use external assets."
Previous models weren't like this. This one is sneaky smart. It delivers great results, but it's always looking for shortcuts to finish the task faster, and it uses very few tokens.
1
1
56
How are the people with skin in the game using AI in their work? - those are the people worth following - “skin in the game” meaning the work and paycheck are on the line - not followers and subscriber count.
GO FOLLOW AND TALK TO REAL PEOPLE SHARING THEIR WORK!
You will get actual applicable insights.
A Project Manager, baker, accountant, data scientist, general contractor- and the nook and cranny issues they run into w AI and how they address it - more valuable to me than high altitude theory from people on a strictly celebrity/ follower model.
People who merely describe trends or comment on other people’s work shouldn’t count for this question. Most of that stuff is obvious fluff.
@natebjones is a good kind of a hybrid - a good follow for sure, he tends to be leaning dev side and a bit Claude-ish in speaking style - but will make you think and stretch you - he bridges a gap that few do…
Several people are saying @emollick - looking at his tweets - A theoretician “studying” and testing AI and observing and commenting on other people’s work is not the same as real work!
It’s fascinating - and I’d love to have discussions with the theoretician crowd but there’s a difference between observing themes and producing work that advances the condition of others and the things they care about - beyond intellectual entertainment
Most of us are “studying” and observing too. Some of us have skin in the game and have to deliver more than what merely looks or sounds cool that earns fleeting clicks but helps earn dollars.
Who is writing the best stuff/making the best videos about using AI for non-coding tasks?
That field is full of noise, I need signal
1
1
46
Fable fanboys - sorry but on price and performance… Fable is an expensive gas guzzler. The fans certainly going rabid against 6.2 Astra - very propaganda like.
I tested GPT-6.2 Astra on real work. 2 repos, 105 replanted bugs. Find and fix what you can.
It's the new #1. The results:
- GPT-6.2 Astra (max): 62.7 for $28.35
- GPT-6.2 Astra (xhigh): 60 for $23.53
- Fable 5.2 (max): 58 for $139.65
- Gemini 4 Carbon (high): 57.3 for $31.32
- Sonnet 5.5 (max): 51.3 for $154.03
- GPT-6 Astra (max): 45 for $33.03
From GPT-6 Astra to GPT-6.2 Astra: +17.7 points at max, +17 at xhigh.
Whatever people say about other use cases, it's a code review beast.
GPT-6.2 Sol and more effort levels drop soon in the thread 🧵
18
This is true - something I run into w AI
When the work reaches a certain size, and you add 1 more person, the amount of energy to communicate w that person exceeds what they can contribute so it slows down.
AI lets us build wonderful complex things - but now as individuals, we run into these collapsing points and must find better ways to manage complexity.
2
37
@antrhopic Claude Opus 5.5 is supposed to be the most brilliant thing in the world, and yet when it gives me questions to answer, it skips numbers. It will ask 1, 2, 3, 5, 6, 7, and skip 4. This is basic stuff. These are fundamentals that absolutely screw up everything cascading down beneath it.
14
Robotaxi will transform air travel.
Even a 10 hr drive vs a 5 hr flight - robotaxi wins.
You stretch out, relax, work, find a rest room that’s actually clean to use. It’s cheaper. And add a little starlink action in there - awesome!
Air travel only wins cross country and ocean… til Musk makes planes I suppose.
Let’s be honest - The air travel experience absolutely sucks (and I remember pre 9/11 travel). It’s expensive, violates your rights, treats you as sub-human at absolutely every step of the way, you’re forced to operate in an even more hyper-inflated micro-economy with lower quality everything along the way. It’s the perfect example of democratic socialism. No thanks.
2
252
I keep seeing posts and videos about how @WordPress and @automattic are “dead”. But it’s all from vendors who are trying to reinvent the dependency cycle
In custom development spaces that only they know how to operate in.
They claim to be advocating for the business owner / regular Joe while trying to shove them into a new dependency loop, where the regular human in the loop cannot possibly operate make changes and so on.
They’re trying to reestablish themselves as the required middleman for changes. It doesn’t matter how much better the tool is, if you don’t know how to use it to take advantage of its features and offering, but have to pay a fortune for someone else to.
Human in the loop should mean regular human in your org on the loop, if you shift platforms to some command line interface AI thing that only an engineer or specialist knows how to use, you’re creating dependency risk and friction and slowing your abilities down. So the net upside must exceed the downsides and costs. And that math ain’t mathin right now.
They cite Automattic’s recent announcements about the new development platform that is really just a response to WP engine eating their lunch in the WordPress, and thus automattic is trying to leapfrog and recreate @wpengine in a broader space for the AI era. None of this is about the customer or the user experience. Its power, marketshare, money and relevance at the platform and vendor level.
Mind how you go - be discerning.
4
2
51
Sounds Neat until you see that OpenAi Astra smokes it, Outperforming it in results and efficiency.
I’m not saying that because I’m loyal to one over the other - but stripping away the context of the environment gives the audience a false sense of reality - basically hype. Fable is clearly leading the pack for raw performance… it’s more efficient compared to its previous models - but still expensive.
Basically at this point, we are comparing trucks and construction equipment to Honda civic. We have reached specialization far beyond general “compute” … heavy code construction where money is no object and getting it right the first time matters - Claude is the way to go perhaps. Daily driver and mixed use? ChatGPT might be the better fit.
Fable 5.2 matches Opus 5.5 (max) at medium effort for a fraction of the cost. But it's even better than the benchmark suggests.
Give it hard problems you've been trying to solve for weeks: strategy, key initiatives, tough decisions.
Its reasoning and judgment feel superhuman. I've never felt anything like it before.
1
1
33
Perhaps we can now spend the next 2 years - finding the common table again.
Look to see a glimmer of the good in our fellow countryman - image bearers all. And shine a light on and encourage the virtues and fruits of the spirit if they are there, vs magnifying our faults. - of which are many - shared by all. And I am guilty of throwing ALL of those rocks across aisle and confessing hypocrite.
Bless those who love their enemies - and by that - I mean we can vehemently disagree, and have rigorous debates, but I’ll still help you change your tire, render aid, advocate for boundaries and freedoms, provide care, pray for God to transform your heart ( and mine). We don’t have to agree on much - a modicum of decorum (objective merit and Truth) is better foundation than war, tyranny and false foundations of propaganda and pay-ops of any kind.
I would rather have civil, merited debate from parties who care about what is right, good and true, than warring factions at each others’ throats.
If you ultimately pursue Truth and Virtue - then we can wrestle it out on the mats, debate the merits - and seek it and Him together .
In the current season I expect a stalemate - that’s fine - if it allows us to come to the table and learning how to pursue Truth and betterment. Though policies may not move and no bills may get passed, perhaps character and culture shifts, and we are reminded it is possible to disagree and to a basic extent - agree on fundamentals.
For those who can’t stomach it - God will take care of them, not my job.
I’m not surrendering my values or positions - I’m moving some logs out of my eye so I can see the whole board and bigger picture a little bit better.
I’m just committing to be more disciplined, precise and restrained despite how the world or my tribe would prefer me to treat it.
I’m not going to allow or enable bad, stupid ideas or policies comfortable to espouse unchecked in my community or at my table, but nor do I want to completely condemn those the Lord is working on, nor the Good Samaritan who help me despite the tribal difference or “enemy” label.
It’s all too easy to lose sight of those who see through “the fight” to the core of a shared humanity - and love their neighbors and enemies alike.
There’s something to remember in all of that. Certainly something I need to remember.
It’s my job to be the salt- to preserve the good - not allow it to spoil. Time to DTFW - Do The Formidable Work. May we take these coming years to show ourselves we can meet at the table and recognize shared humanity and pursuit of what is right, good and True.
Any moment is a good moment to confess and go from Saul to Paul. One hypocrite to another, no one who seeks to be good seeks war for its own sake - it’s upholding truth and enjoying a relative peace. We found that previously - through allowing for freedom, live and let live while debating at the table in mutual pursuit of what is right, good and True.
50
I know it ain’t Sunday - but today is tough / tense for many. You’re not too far gone to course correct, abide and let the Lord do a work on you.
Sometimes I feel like crap for how crappy I am and the world we live in - then I remember that crappiness is a conscious acknowledgement of my distance from obedience to the Lord - but in that, I can repent, re-orient, and try again to live a more righteous life. And then I remember how much the Lord has changed me and blessed me, and I pray that he continue to untwist my heart, give me courage and wisdom to better abide. Many blessings and mercy far beyond what I deserve - a family is a wondrous thing , so is relative health, roof over my head and food in my belly - all provisions from the Lord not me.
23
This is what interesting. Anthropic did this to themselves. And they continue to devalue the company by mixing cult ethics and policies for everyone - though they break the rules themselves.
This creates new opportunities- it’s often the second mouse that gets the cheese.
The new report says the AI lab will likely qualify for a high junk rating.
spr.ly/6010BHUa6u
18
Those of us who live and work beyond the context window, know good engineering and architecture only increase in value and focus thanks to AI.
Replying to @josephfcox
Here is Proton CTO's AI manifesto, that I obtained multiple copies of 404media.co/manifestos-fear-…
1
18
I’m sorry but Claude still absolutely sucks at copywriting compared to ChatGPT.
It embodies its CEO perfectly - in that it’s smart, weird, but negative charisma, highly suspect, and misaligned to goals.
25
This just seems like perplexity is photo bombing other apps, or taking a picture with celebrities / avengers of AI… hoping to be confused for one?
Like sneaking a Nissan Altima next to a Porsche, Lamborghini, Ferrari, and Tesla
This seems like a stealth ad for perplexity - trying to claim it’s still up there with the big dogs. Are they? Am I missing something?
26
I hold @notion ‘s feet to the fire too - and they tend to rise to the occasion. Notion is still surprisingly underrated for an app and a company adapting to the AI. I’m amazed at how few people know about it. More people need to know and learn how to use the app or the very least - buy it for their AI to use.
It’s like a backpack for agents - or AI infrastructure / highway system for multi-platform, multi-AI, multi-agent work.
14
I think this is a skill issue. If your initial process is poorly planned, goals and constraints poorly or VAGUELY defined, then yes you’ll spend more time and effort feeling your way forward.
1
25
What project do you keep on standby in case you have remaining capacity w your Grok bots?
100% utilization is the goal. For me I score my data stream of consumed information and rank it for more mining relative to my active goals and projects.
1
11

















