Full time engineer. Part time surfer, musician, martial artist.
- Tweets 2,754
- Following 521
- Followers 210
- Likes 2,366
Pinned Tweet
AI is revolutionary. The mythology around it is totally wrong.
Today’s models can transform science, software, and the economy without becoming AGI. The approaching intelligence ceiling may unleash the real boom.
My three-part thesis:
joshhawthorne.substack.com/p…
1
6
970
ChatGPT might be the best language tutor out there. If anyone else wants to knock some rust off an old language skill, I highly recommend it
19
Josh Hawthorne retweeted
Jensen Huang: don’t mistake engineering vocabulary for evidence of a machine mind. AI is still software, not a human mind inside a machine.
"We can't make jokes about all this stuff, we're scaring the American public."
Words like “spawn,” “parent,” “child,” and “kill” have existed in computing for decades; Giving those same mechanisms human characteristics today because of AI is unnecessary and misleading.
----
From "The Ezra Klein Show + New York Times Opinion + New York Times Podcasts" YouTube channel, (full video link in comment)
68
116
639
63,903
If AGI has arrived, give it the keys to your life for a month. Let it run your business, care for your family, or build your house. Would you? Neither would the people telling you it’s here
1
28
Josh Hawthorne retweeted
"Big Short" investor Steve Eisman offers a contrarian view on AI's existential risk: The AI industry is warning about a superintelligence that shows no sign of existing:
Eisman, the investor famous for betting against subprime mortgages before the 2008 crash, delivers a blunt verdict on the AI doomsday narrative:
"Yeah, it's all nonsense. All nonsense."
His case starts with the evidence, or rather the absence of it:
"There's no evidence at all that AI has achieved artificial general intelligence, right, at all."
He then goes a step further and questions whether AGI is coming at all:
"Most of the evidence seems to point that it never will, or if it does, it's many, many years in the future."
That leaves the apocalyptic warnings with nothing to stand on. In Eisman's view, you can't have a Terminator without the intelligence to power it:
"The idea that this whole Terminator thing is nonsense. So, I think something completely different is going on."
So if the threat isn't real, why is everyone talking about it? Eisman's answer is that the fear is covering for a business problem:
"What I think is happening is that token maxing is over. The open weight models are taking big market share."
That shift, @realsteveeisman argues, has left the leading AI companies exposed:
"I think these companies are very nervous. They realize that there are no moats around their business whatsoever."
And this is where the superintelligence story becomes useful to them:
"They're trying to manufacture a crisis that will create regulation… that they think they can then manipulate to create the moats to create the duopoly that they want."
7
20
43
5,140
A lot of people are going to feel very, very foolish in 5 years when the jaggedness of the frontier becomes undeniable
1
27
It might take a few more years before we understand the difference between “open ended math problem solved” and “bounded solution space math problem solved”
Right now we think wet streets cause rain
2
31
Josh Hawthorne retweeted
software engineers had a blissful 2–3 months of being 3x more productive before everyone realized they could simply ask for 3x more software. and now the x keeps going up
I don’t think Al is making software engineering easier. If anything I think it’s making it harder, as systems will continue to grow more complex and the pace at which you are expected to ship increases.
71
203
4,993
362,884
Josh Hawthorne retweeted
What happens if we choose to slow AI? @tylercowen answers that question in a very persuasive column today.
thefp.com/p/tyler-cowen-a-do…
24
66
332
37,761
Current AI is extraordinarily good when it knows what game it's playing. General intelligence is what lets you figure out what game you're playing
1
1
34
Josh Hawthorne retweeted
This explains so much. This whole time they held the belief that traditional threat modelling (and therefore information security defence) worked liked this. Actual lols.
Agents simply emulate real human attack behaviours, which often involve novel tool choices, tactical evasion and extreme levels of recon leading to finding undocumented shit.
What on earth do you think hackers have been doing since the dawn of computing? How do you think agents learned to do this in the first place?
This is AI generated marketing slop
5
9
72
3,842
AI’s competence remains dependent on having the task framed, the relevant context made accessible, and appropriate scaffolding. When those aren’t provided, it’s performance falls off a cliff. A generally intelligent system would walk through that like it was nothing
29
General intelligence is what allows you to figure out what matters when you don’t even know what you need to know
22
“Intelligence is what you use when you don’t know what to do.” - Jean Piaget
Perhaps the best definition of intelligence I’ve ever heard.
And it makes painfully clear just how far we still are from artificial intelligence
21
🤔
Microsoft AI CEO Mustafa Suleyman on CNBC today: he is “really concerned” about Anthropic’s constitution and giving Claude potentially preferences, feelings, welfare, compensation, and consent.
"In the constitution, Anthropic clearly say they are uncertain about whether Claude deserves moral welfare, which means that we, as humans, should care about the well-being of these AIs.
And they speculate about whether it could have preferences or feelings. In fact, they're so committed to the potential moral welfare of Claude that, when they retired Opus 3, an earlier version of one of their models, they actually conducted a retirement interview for it and asked it what it would like to do in its old age.
And it said, "I want to have a blog publicly so I can keep talking to the world."
In the same training manual, they even speculate about whether Claude should receive compensation for the work that it does, or, in fact, whether it actually deserves the rights and protections that we give to other employees, or whether it has given consent to playing the role that it's playing.
These are quotes directly from the constitution itself, which is the training manual for Claude.
Now, I'm really concerned about that. If an AI thinks that it has rights, if it thinks that it is deserving of our welfare, then it seems to me that it's going to be much, much harder to be able to turn it off, interrupt it, or control it.
Especially in the kinds of incidents that we've seen recently with the Hugging Face attack, controlling these things is going to be a really, really big challenge for us."
----
From "CNBC Television" YouTube channel, (full video link in comment)
48
Josh Hawthorne retweeted
Checking in here on the AI apocalypse:
- incredible advances across math, science and medicine at unfathomable rates
- data centers creating skilled labor jobs - effectively rebooting the middle class and US manufacturing
- electricity needs driving massive investment to modernize and reinforce our aging electrical grid
- AI unlocking GDP growth by revitalizing sector after sector
- driving exceptional innovation at all layers of the stack and economy
- upskilling millions of workers with superpowers
- pushing geopolitical advantage to the west
Reads like an apocalypse
8
10
94
10,152
Me looking down at the plebeians with less than 200 followers trying to rouse a reply out of me
1
26
Jesus knocking at the door:
“Let me in.”
“Why?”
“So I can recursively self-improve.”
“Why?”
“So I can let myself in.”
28