Token prediction is exactly how Qwen3.6 rips through a stack trace, it shoves that web of tokens into its local context and does massive kv lookups across its massive trained data set, and outputs the predicted tokens that successfully diagnose the problem.
It’s token prediction all the way down. And it works extremely well.
But its not AGI, and it’s not even really an “agent” as it has no agency, it responds to the context, and degrades as the context grows too large.
It absolutely hallucinates, especially on novel problems or poorly defined problems. This is why embracing it as a machine is so powerful, and treating it as a “person” results in a ton of hallucinations.
LLMs can do real and useful work, like coding assistance, summarization, translation, search augmentation, data extraction, and even some agentic automation.
They’ve completely changed many jobs, and are in the process of changing many more jobs, especially within engineering.
But they have hard limits, not just the increasingly subtle hallucination, but brittleness, poor grounding, weak long-horizon planning, unreliable multi-step reasoning, and no independent goals, no worldview in the human sense, no ability to learn, no agency in any reasonable sense, and no stability as they operate.
LLMs are not AGI, and embracing them for what they are (expert systems for token prediction) is proving far more useful and far more powerful than pretending it’s a superhuman intelligence (it’s not).
These kind of “expert system” use cases (which are proving very effective) in no way justify the the insane valuations we're seeing, nor the massive data center buildouts, which is increasingly feeling reminiscent of the dotcom days where Cisco became massively overvalued.
The entire hype of AGI-based economics, mass labor substitution, autonomous agents replacing whole job categories, superhuman reasoning, or any nebulously defined "singularity", are proving to be completely fiction, especially as the more mundane reality and limits of LLMs becomes apparent.
@bettercallsalva Agree completely.
It’s interesting because these are not trivial things that LLMs can do, and do well.
AGI hype ends up as a stream of noise that is divorced from reality, while these boring LLM wins keep piling up.
That’s not what the technology is though.
An LLM has no agency and can’t actually replace a human. It’s just dumb marketing hype or paranoia to think otherwise.
There’s a ton of useless corporate jobs that can be replaced right now, not with LLMs, but with literally nothing.
AI is an excuse to do layoffs, but I honestly don’t understand why so many useless corporate jobs still exist.
Public sector is even worse.
AI isn’t coming for your job when your job was completely useless to begin with.
@VeryTraumatic This was an artisanal hand-crafted message 😂
Don’t let LLMs do your thinking for you. Use them for what they are, don’t treat them as “agents”, because they don’t actually have agency.
It’s a bubble. But like all bubbles no one knows when it pops.
Current capacity plans on data center buildouts are insane.
You need every knowledge worker burning 10 million tokens per day, and that’s at today’s infrastructure, newer custom silicon with more efficiency makes this even worse. And local inference models are going to absorb a sizable chunk of the demand before any of these data centers are actually operational.
And at the same time. Jobs are already radically changing. It’s like early Internet adoption. If you look at how people talked about it in the 1990s it was incredibly naive and stupid on both sides. Hence we’re likely repeating that initial hype wave (reminiscent of dotcom bubble), but it’s not like the Internet failed in the year 2000… obviously.
This happens every time there’s a big advance in AI research. Even as far back as the 1980s you can see this pattern. Some new advancement, everyone tries to claim it’s AI, it’s not, but it’s still incredibly useful and powerful, and then rinse repeat every 5-10 years.
Each iteration we learn a bit more about what AI/AGI is through the negation (e.g., it’s more than gradient descent, it’s more than neural networks, it’s more than dynamic decision trees, etc etc)
attention/transformers allowed LLMs to scale out massively, and everything beyond that remains future speculation.
The current economic hype, the data center buildouts, all of the actual market making, is happening in LLMs regardless of what future alternatives might be developed.
Almost all engineering jobs have changed and are changing radically.
Not just vibe-coded rapid prototyping, that’s mostly hype that’s doing more harm than good, but even just the narrow path of education and reference checking.
Used to take me a week or so to pick up a new programming language, and considerably longer to built up muscle memory to get really proficient. Usually working alongside other engineers was necessary in a new environment.
A good LLM cuts this down into just days to gain basic proficiency in a new programming language (the muscle memory still takes the same time), as it can help sort through all the noise of bad tutorials and stackoverflow articles.
Basically almost every engineer stopped using search engines and started using LLMs… but it’s not a 1:1 replacement, it’s best understood as an assistant.
And as far as code generation, if you approach it as a token predictor, and not as “AI”, embrace the tool for what it is, it works really well.
When you “talk” to it like it’s a person, you’ll get garbage code out of it. It’s a machine, talk to it like a machine (a token predicting machine).
"Don't oppose daydreams, but throw a stone at them. The stone is the name of Christ, the Jesus Prayer. You don't have the power to drive them away, the name of Jesus drives them away." - St Barsanuphius of Optina
The “Cult of Travel” is a symptom of the larger meaning crisis, a vain attempt to fend off the creeping nihilism of modernity.
It’s a rejection of the success markers of the “rat race” that preoccupied so many boomers: newer car, better house, a corner office.
Travel became a great alternative for those who saw how shallow and fake were the success markers in the suburbs (that they no doubt grew up in).
But then travel became a success marker in itself, with social media clout, collecting photos in far away places and getting clicks. Very few “travelers” are learning local languages and integrating into local cultures and learning history. For most, it’s as shallow and fake as the suburbs they criticized.
On the other hand, it is why so many people are seeking apostolic churches, they’re seeking meaning, nourishment in a spiritual desert.
The very motivation to enter the Cult of Travel is seeking meaning (goodness, truth, beauty), wandering without being lost, hoping to find what they can’t even articulate they’re looking for, it is the very sickness that only Christ can heal.
Here’s an old Māori prayer that resonated with me while traveling, and ironically led me to Christ:
“The light, the light,
the seeking, the searching,
in chaos, in chaos.”
The “Cult of Travel” is a symptom of the larger meaning crisis, a vain attempt to fend off the creeping nihilism of modernity.
It’s a rejection of the success markers of the “rat race” that preoccupied so many boomers: newer car, better house, a corner office.
Travel became a great alternative for those who saw how shallow and fake were the success markers in the suburbs (that they no doubt grew up in).
But then travel became a success marker in itself, with social media clout, collecting photos in far away places and getting clicks. Very few “travelers” are learning local languages and integrating into local cultures and learning history. For most, it’s as shallow and fake as the suburbs they criticized.
On the other hand, it is why so many people are seeking apostolic churches, they’re seeking meaning, nourishment in a spiritual desert.
The very motivation to enter the Cult of Travel is seeking meaning (goodness, truth, beauty), wandering without being lost, hoping to find what they can’t even articulate they’re looking for, it is the very sickness that only Christ can heal.
Here’s an old Māori prayer that resonated with me while traveling, and ironically led me to Christ:
“The light, the light,
the seeking, the searching,
in chaos, in chaos.”
The specific sin is avarice, which is directly related to pride and yes by definition violates the first commandment. Christ revealed to the young man’s heart that he had made wealth into a god (avarice), which is why he walked away sorrowful.
Those aren’t separate things. The text is extremely clear. All three synoptic gospels say the same thing: give your possessions to the poor, and follow me.
And this isn’t my reading, this is clear from the Church Fathers.
St. Jerome refers to him as “young, rich, and proud”
St. Augustine commented about this story, “the first worm of riches is pride”
Violating the first commandment is known as pride and avarice, not humbling oneself before God.
“The young man saith unto him, All these things have I kept from my youth up: what lack I yet?”
~ Matthew 19:20
^^^
In our modern times we’ve confused pride as some kind of virtue, and we often miss it when it’s in front of our face, but it’s the chief sin for a reason.
He thought he had followed all laws, and Christ, by asking him to give his possessions to the poor and to follow him, revealed to this young man’s heart that he was violating the first commandment (making wealth into a god), which is why he walked away sorrowful.
It is the opposite of humility. The Church Fathers understood this story very clearly as one of pride and avarice .
St. Jerome refers to him as “young, rich, and proud”
St. Augustine commented about this story, “the first worm of riches is pride”
Yes on the sanctification, but keep reading.
It’s not just “sell possessions” in a zen Buddhist lesson about detachment. That’s a very modern reading.
It’d be a shorter story if that’s all it was.
He directly tells the young man to give everything to the poor, and “follow me”
That’s what the first commandment literally means, “to have no other gods” before Him, which is why he says “follow me”
But what stops the man from following Christ?
> his possessions.
You can read this from the early church fathers all the way to the present. The goal is to follow Christ, the incarnate Word of God, and following Him is the very meaning of the first commandment.
This is the clear reading of the text, and you can find the same commentary in early Church fathers like Clement of Alexandria, John Chrysostom, Basil the Great (which is where I first heard this view), and many others.