If you read nothing else today, read Matt Shumer on AI

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Pp who works with AI agents - np above is on point. The author is on the inside and frankly doing a public service

Claude skills and cowork has rocked a lot of companies this January. Everyone is sprinting to adopt - it’s not hype or futurist predictions anymore.

This post is such transparent marketing hype. This is all a desperate attempt to make AI happen as these overvalued companies are hemorrhaging money in this silly endeavor.

I’m the np above. Look, I get it, I’m skeptical and lean towards being a Luddite. And AI can do dumb things. One of my co-workers described it as like working with an eager intern who needs to be reined in sometimes. But the changes are real. The improvements in its quality are exponential.

I don’t really know what this means for the future of work, especially for my kids who are still in high school, but this isn’t smoke blowing. Disruptive change is coming.

Ok but what ARE the changes? What are the exponential improvements? Where can I look and see for myself something completed with AI that is really mind-blowing? People keep talking about AI doing things but provide no evidence of AI actually doing the thing. This is not coming from a place of skepticism; it’s just a basic question that no one seems able to answer.

Again, if you read the article, it explains it well. Most people who "use" AI are using older models that are still a bit buggy. There's a new generation, as in the last month, that doesn't just do stuff 80% of the way -- it's 100% now. It's perfect. These two are GPT-5.3 Codex from OpenAI, and Opus 4.6 from Anthropic (the makers of Claude, one of the main competitors to ChatGPT).

The leap, according to the article, is "I describe what I want built, in plain English, and it just... appears. Not a rough draft I need to fix. The finished thing. I tell the AI what I want, walk away from my computer for four hours, and come back to find the work done. Done well, done better than I would have done it myself, with no corrections needed. A couple of months ago, I was going back and forth with the AI, guiding it, making edits. Now I just describe the outcome and leave."

He then goes on to say you can try it for yourself, but you have to pay the $20 and be sure to be using the latest version. "Sign up for the paid version of Claude or ChatGPT. It's $20 a month. But two things matter right away. First: make sure you're using the best model available, not just the default. These apps often default to a faster, dumber model. Dig into the settings or the model picker and select the most capable option. Right now that's GPT-5.2 on ChatGPT or Claude Opus 4.6 on Claude."

So, your question, "Where can I look and see for myself." It's right there. He tells you how to do it.

We are going in circles. I DO use the paid versions of Claude and ChatGPT. I do NOT see this “leap” that the article (by someone with a vested interest in hyping his product) discusses.

WHERE are these amazing things being built by AI without user input? Where are they in YOUR work?

So, you must not be very good at prompting. He covers this, too. My guess is you ask it questions, treating it like it's Google.

As I said, I'm a writer. I can see the dramatic improvements versus the output only a year ago.

Ok. So give us an example of something complete that the AI has written for you. Or better yet, go ahead and use your superior prompting skills to generate something you would use for your work, since it requires virtually no effort from you.

I'm literally doing that right now, producing thought leadership to make a case for cosourcing certain professional services. That's all I can say without revealing too much of my identity. I was having trouble making sense of a couple of interviews I had with executives earlier this week, so I uploaded my notes from those interviews to an AI and asked it to generate a 1,000-word article that incorporates them along with the findings of a couple of recent survey reports and at the moment I'm playing around with whether or not I want the output suitable for a LinkedIn post or an article for our website. This involves a series of prompts. But the immediate output is heads and shoulders better than anything from a year ago and was generated in seconds versus something that might have taken me several hours to do, especially since I was having some trouble wrapping my head around the angle for this one.

OMG, you're "producing thought leadership" by feeding interview notes to an AI to write an article for LinkedIn? Life is satire.

What? We all do this to shortcut the drafting process. Why aren't you?
Lawyer here.

I don't "produce thought leadership" on LinkedIn, I do real work.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Pp who works with AI agents - np above is on point. The author is on the inside and frankly doing a public service

Claude skills and cowork has rocked a lot of companies this January. Everyone is sprinting to adopt - it’s not hype or futurist predictions anymore.

This post is such transparent marketing hype. This is all a desperate attempt to make AI happen as these overvalued companies are hemorrhaging money in this silly endeavor.

I’m the np above. Look, I get it, I’m skeptical and lean towards being a Luddite. And AI can do dumb things. One of my co-workers described it as like working with an eager intern who needs to be reined in sometimes. But the changes are real. The improvements in its quality are exponential.

I don’t really know what this means for the future of work, especially for my kids who are still in high school, but this isn’t smoke blowing. Disruptive change is coming.

Ok but what ARE the changes? What are the exponential improvements? Where can I look and see for myself something completed with AI that is really mind-blowing? People keep talking about AI doing things but provide no evidence of AI actually doing the thing. This is not coming from a place of skepticism; it’s just a basic question that no one seems able to answer.

Again, if you read the article, it explains it well. Most people who "use" AI are using older models that are still a bit buggy. There's a new generation, as in the last month, that doesn't just do stuff 80% of the way -- it's 100% now. It's perfect. These two are GPT-5.3 Codex from OpenAI, and Opus 4.6 from Anthropic (the makers of Claude, one of the main competitors to ChatGPT).

The leap, according to the article, is "I describe what I want built, in plain English, and it just... appears. Not a rough draft I need to fix. The finished thing. I tell the AI what I want, walk away from my computer for four hours, and come back to find the work done. Done well, done better than I would have done it myself, with no corrections needed. A couple of months ago, I was going back and forth with the AI, guiding it, making edits. Now I just describe the outcome and leave."

He then goes on to say you can try it for yourself, but you have to pay the $20 and be sure to be using the latest version. "Sign up for the paid version of Claude or ChatGPT. It's $20 a month. But two things matter right away. First: make sure you're using the best model available, not just the default. These apps often default to a faster, dumber model. Dig into the settings or the model picker and select the most capable option. Right now that's GPT-5.2 on ChatGPT or Claude Opus 4.6 on Claude."

So, your question, "Where can I look and see for myself." It's right there. He tells you how to do it.

We are going in circles. I DO use the paid versions of Claude and ChatGPT. I do NOT see this “leap” that the article (by someone with a vested interest in hyping his product) discusses.

WHERE are these amazing things being built by AI without user input? Where are they in YOUR work?

So, you must not be very good at prompting. He covers this, too. My guess is you ask it questions, treating it like it's Google.

As I said, I'm a writer. I can see the dramatic improvements versus the output only a year ago.

Ok. So give us an example of something complete that the AI has written for you. Or better yet, go ahead and use your superior prompting skills to generate something you would use for your work, since it requires virtually no effort from you.

I'm literally doing that right now, producing thought leadership to make a case for cosourcing certain professional services. That's all I can say without revealing too much of my identity. I was having trouble making sense of a couple of interviews I had with executives earlier this week, so I uploaded my notes from those interviews to an AI and asked it to generate a 1,000-word article that incorporates them along with the findings of a couple of recent survey reports and at the moment I'm playing around with whether or not I want the output suitable for a LinkedIn post or an article for our website. This involves a series of prompts. But the immediate output is heads and shoulders better than anything from a year ago and was generated in seconds versus something that might have taken me several hours to do, especially since I was having some trouble wrapping my head around the angle for this one.

OMG, you're "producing thought leadership" by feeding interview notes to an AI to write an article for LinkedIn? Life is satire.

What? We all do this to shortcut the drafting process. Why aren't you?
Lawyer here.

I don't "produce thought leadership" on LinkedIn, I do real work.

Lol, in a few years, people will assign their personal bots to read all the Thought Leadership on LinkedIn and bring back the top 5 one-line bullets they need to know.

Yes they can do that today but some people still choose what they digest.

[Anonymous]

Anonymous wrote:The people on this site are the most pessimistic about AI I have ever met. I think part of it is because many in the DMV who frequent this site are extremely well paid and have relative safe jobs and they do not want the gravy train to leave.

We can't predict the future of AI. But to dismiss anyone who speaks about its potential tells me you are just as terrified about your future standard of living should some of the predictions about AI and white collar jobs turn out to be true.

I’m not pessimistic about AI. I just don’t think this article backed up its own claims.

The author says AI built him a perfect app instantly. Ok, where is it? What does it do? Can I, the reader, test it out and see if his claim holds?

I would like to say this is critical thinking but I don’t think it even rises to that. If someone says “I built this cool thing” the most natural response is “what is it?”

[Anonymous]

Anonymous wrote:

Anonymous wrote:The people on this site are the most pessimistic about AI I have ever met. I think part of it is because many in the DMV who frequent this site are extremely well paid and have relative safe jobs and they do not want the gravy train to leave.

We can't predict the future of AI. But to dismiss anyone who speaks about its potential tells me you are just as terrified about your future standard of living should some of the predictions about AI and white collar jobs turn out to be true.

I’m not pessimistic about AI. I just don’t think this article backed up its own claims.

The author says AI built him a perfect app instantly. Ok, where is it? What does it do? Can I, the reader, test it out and see if his claim holds?

I would like to say this is critical thinking but I don’t think it even rises to that. If someone says “I built this cool thing” the most natural response is “what is it?”

It’s pretty transparently false, but the AI salespeople are a cult, so…

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Pp who works with AI agents - np above is on point. The author is on the inside and frankly doing a public service

Claude skills and cowork has rocked a lot of companies this January. Everyone is sprinting to adopt - it’s not hype or futurist predictions anymore.

This post is such transparent marketing hype. This is all a desperate attempt to make AI happen as these overvalued companies are hemorrhaging money in this silly endeavor.

I’m the np above. Look, I get it, I’m skeptical and lean towards being a Luddite. And AI can do dumb things. One of my co-workers described it as like working with an eager intern who needs to be reined in sometimes. But the changes are real. The improvements in its quality are exponential.

I don’t really know what this means for the future of work, especially for my kids who are still in high school, but this isn’t smoke blowing. Disruptive change is coming.

Ok but what ARE the changes? What are the exponential improvements? Where can I look and see for myself something completed with AI that is really mind-blowing? People keep talking about AI doing things but provide no evidence of AI actually doing the thing. This is not coming from a place of skepticism; it’s just a basic question that no one seems able to answer.

Again, if you read the article, it explains it well. Most people who "use" AI are using older models that are still a bit buggy. There's a new generation, as in the last month, that doesn't just do stuff 80% of the way -- it's 100% now. It's perfect. These two are GPT-5.3 Codex from OpenAI, and Opus 4.6 from Anthropic (the makers of Claude, one of the main competitors to ChatGPT).

The leap, according to the article, is "I describe what I want built, in plain English, and it just... appears. Not a rough draft I need to fix. The finished thing. I tell the AI what I want, walk away from my computer for four hours, and come back to find the work done. Done well, done better than I would have done it myself, with no corrections needed. A couple of months ago, I was going back and forth with the AI, guiding it, making edits. Now I just describe the outcome and leave."

He then goes on to say you can try it for yourself, but you have to pay the $20 and be sure to be using the latest version. "Sign up for the paid version of Claude or ChatGPT. It's $20 a month. But two things matter right away. First: make sure you're using the best model available, not just the default. These apps often default to a faster, dumber model. Dig into the settings or the model picker and select the most capable option. Right now that's GPT-5.2 on ChatGPT or Claude Opus 4.6 on Claude."

So, your question, "Where can I look and see for myself." It's right there. He tells you how to do it.

NP. I know plenty of people paying for the best models who are rolling their eyes at this article. The article itself is marketing slop.

[Anonymous]

Anonymous wrote:
This is alarming, but not alarmist. It is making the rounds this week. It’s a must-read to understand what it about to happen. Not in 10 years, more likely in the next one-to-two.

Maybe I will retrain to be a plumber or something.


https://shumer.dev/something-big-is-happening?fbclid=IwZnRzaAP6pfdleHRuA2FlbQIxMQBzcnRjBmFwcF9pZAo2NjI4NTY4Mzc5AAEeg3Oij6mY1B_GCLqt_RggOSduVkePMwV6HKnMwZemWFZSzQFbaN3FPSKRUgI_aem_MRW-NPtQq1TktuKMS-kdUg

Now read an article from someone who is good at spotting fakery and who isn’t selling a product:

https://www.theguardian.com/us-news/ng-interactive/2026/jan/18/tech-ai-bubble-burst-reverse-centaur

[Anonymous]

I am sure plenty of you own an Index fund correct? You know that your favorite brokerage such as Vanguard, Fidelity etc will sooner or later use AI in their decision making correct?

So be skeptical of AI all you want, the folks managing your money will soon be using AI.

[Anonymous]

Anonymous wrote:I am sure plenty of you own an Index fund correct? You know that your favorite brokerage such as Vanguard, Fidelity etc will sooner or later use AI in their decision making correct?

So be skeptical of AI all you want, the folks managing your money will soon be using AI.

Do you know what an index fund is? I'm pretty sure you don't need AI to structure a fund that follows a PUBLISHED index

[Anonymous]

Anonymous wrote:

Anonymous wrote:I am sure plenty of you own an Index fund correct? You know that your favorite brokerage such as Vanguard, Fidelity etc will sooner or later use AI in their decision making correct?

So be skeptical of AI all you want, the folks managing your money will soon be using AI.

Do you know what an index fund is? I'm pretty sure you don't need AI to structure a fund that follows a PUBLISHED index

I think what PP is saying is that people are being skeptical while the stocks selected in their retirement funds are probably going to be selected using AI.

So people are saying AI can't do this can't do that, but it surely can/will select which stocks go into their index funds.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:I am sure plenty of you own an Index fund correct? You know that your favorite brokerage such as Vanguard, Fidelity etc will sooner or later use AI in their decision making correct?

So be skeptical of AI all you want, the folks managing your money will soon be using AI.

Do you know what an index fund is? I'm pretty sure you don't need AI to structure a fund that follows a PUBLISHED index

I think what PP is saying is that people are being skeptical while the stocks selected in their retirement funds are probably going to be selected using AI.

So people are saying AI can't do this can't do that, but it surely can/will select which stocks go into their index funds.

That’s not how an index fund works. Please look up “index funds” on Google.

[Anonymous]

Anonymous wrote:I also think teachers and professors are going to also go away, there is no need to have them when AI can be personalized to cater to everyone's style and be more effective.
Teachers are merely doing attendance and chaperone work organizing papers, this is a big change i am not sure they are ready for.

You are right, I am not. But then again, COVID showed that teaching Kindergarten with bots isn’t really what anyone wants.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:If things are really as good/bad as he says they are then I don’t see what anyone can do.

I do agree that telling your kids to focus on learning/adapting as a skill vs particular subject matters or jobs makes sense but if all he needs to do is tell the AI “build me an app that does x y and z” then it’s kind of stupid to tell me to spend an hour a day “practicing” with Claude.

It’s very hard to tell how much of AI is inevitable and how much people just want it to be inevitable, but if it is inevitable at the level he is talking about then his advice is basically just sticking a finger in the dike and waiting for the economy to implode.

I am also really curious where these law firms expect to find senior partners and if AI replaces all the junior associates.

This is what I’ve been saying. How will you have senior developers if you never have junior developers?

It really just means that the entry level job will be a different kind of job at a different step in the work flow.

Yes, you will need to have basic coding and vibe coding schools by high school. This level used to be junior developer, but not anymore. You are expected to be at mid-level or higher upon college graduation.

[Anonymous]

Anonymous wrote:

Anonymous wrote:I am sure plenty of you own an Index fund correct? You know that your favorite brokerage such as Vanguard, Fidelity etc will sooner or later use AI in their decision making correct?

So be skeptical of AI all you want, the folks managing your money will soon be using AI.

Do you know what an index fund is? I'm pretty sure you don't need AI to structure a fund that follows a PUBLISHED index

See what happens when people use their brains!

Also computerized trading may not be done with an LLM but the financial incentives are already there to capture all the minute advantages with traditional coding. How does it make my financial life different if Vanguard pays a set of coders less (or pays fewer coders) to produce their next gen software? If they can get a bot to help me calculate my basis from before they had to keep mandatory records for life, then cool. They don't have that now.

People don't know how to value white collar work very well. So it's not clear what will happen with the allegedly increased productivity from AI.

I think the Tech Bros are scared because a lot of their business models produce highly-optional digital services and content. So vibe-coding creates more competition for them. If I can make my own mini-Facebook or Twitter for me and my own peeps, I don't need theirs.

[Anonymous]

Anonymous wrote:Can AI predict the next stock crash? Can it solve world hunger? Reduce vehicular deaths? Stop tyrannical dictators? Slow climate change?

I use the free version, but it hasn’t improved my life or my job output. It could not regurgitate simple numbers I fed to it, line by line. Maybe in six months?

Until it can shovel snowcrete it means nothing to me.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Vibe coding....it gets you 80% of the way there. Some folks think that's good enough/don't know to check, others see this as full employment to fix. And this is only going to keep getting better.

I can see you didn't read the article. This is addressed in the beginning. It's not getting you 80% there anymore. It's 100%. And perfect. That's the point he's making. The new models are perfect.

They're perfect the way Phillippine workers are really helping to drive the automated cars perfect.

I appreciated this. Also because a lot of the training originally came from those "click on every streetlight" captchas.