techblueberry

Maybe too many people don't see it this way an it ends up backfiring, but I actually think it's important to "humanize" 'bad' people and show their 'redeeming' qualities -

I think it's important to inoculate yourself against presentation. Like, sure, now we know that Elizabeth Holmes was evil, an we can be binary about it - but would it have easier to be more critical of her rise if we could look past her apparent positive qualities?

This study is fascinating and the one data point I would immediately point out is 86% of Anthropic staff thing Claude can write emails today and the extra 14% think it will happen next year. Which is - weird? What do those 14% think is missing now they what be missing next year?

‘ The bill would let the government block a model’s release or use only if it poses a foreseeable, imminent risk of more than $1 billion in damages, or of killing or seriously injuring more than 50 people.’

The government is not allowed to block a model release with foreseeable imminent risk of killing 49 people?

It says imminent risk, is it 50 people per day? Presumably the model isn’t just one shot killing 49 people it’s recurring right?

‘Growth has slowed" is technically true but misleading. Q2 to Q3 still went from $11.5bn to $16.8bn. And at those levels, it’s not wildly unexpected that it has to slow down.’

The growth slowdown is decidedly unexpected. The market priced in a $2T valuation that assumed it wouldn’t slow down.

https://www.reuters.com/business/finance/investors-temper-an...

Great! Amazon could use some competition.

I think a controversial or complicated question about appreciation is should the narrative around it matter. I watch old movies, partially because I prefer them, but partially because I think real depth of emotion and sustainable meaning and happiness is about having a rich narrative about the world you habit. I watch a movie and connect it to the history I read about what is happening at the time. What was happening in hollywood, who the actors are.

There can be important qualifications to this -> everyone's narrative is different. I think the gotcha around "people can't tell the difference between expensive and cheap wine" is double-sided. If you have the money, telling your self a story about wine that makes you appreciate it more probably enhances your life. If you don't have money - the opposite narrative is also worthwhile.

So, I'm not necessarily saying there aren't benefits to this, and this is the direction the world is going with like instagram reels and what not. But trying to boil the world down to which contextless dopamine trigger is most optimal. I'm not sure I think that leads to a happy life.

And in reverse - why not disclose how it's written? People are different, there are folks whose experience would be enhanced knowing it was written by AI.

This is not strictly speaking true though. I don’t want programmers to turn vague instructions into code. I want programmers to turn vague instructions into context and context into code, often gathering context from sources that aren’t easily obtained.

As a manager I often don’t have enough information to prompt an LLM. That’s what the engineer is for.

Based on what? I think it’s important to take a critical lens. There are lots of pros and cons to DEI so I’d love to dig into the details of your assessment. See if I can update my priors. What evidence have you gathered so far?

If people don’t use muscle power for work, and people don’t think for work, what are people going to fi?

Maybe the new scarce resource will be meaning, and we’ll kill each other for that

It will be curious how anti regulation all these execs are when they start losing money.

> Second, there is no link between bot traffic and Russian disinformation. The Cloudflare data being held up as bot traffic is mainly search engine crawlers harvesting data.

Two things can be true, Cloudflare's graph can be mostly search engine crawlers, and other traffic could be affected.

I’m starting to think we should regulate Anthropic out of existence.

‘They will ask "I have this photo and I want to manually edit it" and it will spin up a Photoshop clone’

This is ridiculous but also I think wrong? ChatGPT is starting to show how dynamic UI will live in the app.

I think for sanity reasons, predictable but flexible applets > custom software for all. I don’t want to learn a new UI for editing images every time I want to add a gradient.

This has very capitalism is everything I like and socialism is everything I don’t like vibes to it.

It’s good to resist excessive labeling, in which case, it FOSS isn’t at least someone disruptive to capitalism-maxxxing. What are we even doing here.

The term “six digits” as an aspirational salary goes at least as far back and 1959, but then it was more like “millionaire”. Still in the 80’s when it became popular as successful upper-middle class professional - 100k in 1983 would be $336,330 today.

$100k today would be 29,712 in 1983 when the median income was $24,580.

I don’t think you have to let the AI companies set the rules. In fact you probably shouldn’t!

“Brad Smith said there needed to be global safety rules in place, including a legal obligation to tell a "victim" if their system is breached by a rogue AI agent…”

Silicon Valley has unprecedented ability to control the global narrative at this point. Everyone loves censorship until it affects them personally.

Reading this post I’m realizing the Superintelligence rebrand was whispered into trumps ear by the industry.

https://youtu.be/5fHYZSoW4GM?is=QjwMtSfM_WJAu9mI

Maybe underneath it all we really do want the same thing…

I don’t entirely disagree with the premise persay - but then I think this warps the incentives, we’ll spend more and more on tokens because there will be more and more forms. We’ll add microtransactions to every form AI parses.

It all seems like an improvement in this world, but wait until the enshittifiers figure out all the ways they can abuse this process. They can add new micro taxes you won’t even know about!

As an SRE, I think this question comes up which is like would you ever trust AI to manage your infrastructure, and I think capability wise maybe, but long term I think it’s going to lead to some wacky incentives/interaxtions.

I’m not the person who reads every line of code, but we are in a golden age of AI right now where “I don’t have to understand anything I can just let the AI do it all” feels like a possible future.

In America anyways, we send the “responsible politicians” to hell every 2-6 years. People vote for this.

Feels like 50/50 Anthropic nor OpenAI exist as independent companies in five years. I remember having this debate that someone was incensed I thought AI might not produce a FAANG company, but people don’t tend to buy pure technology.

In a weird way, looking at the existing companies, it seems “obvious” their struggle will be real. Google is an advertising company, so is Facebook. Most of apples competitors, and Apple itself sell other products. And it’s funny they’re building phones when I think Apple itself is starting to realize the iPhone’s time as a gazillion dollar maker is coming to an end.

Turns out making trillions of dollars is hard.

The soul is whatever story the buyer puts into it. A few years ago, people attributed soul to NFTs, but I think in the long run, humans will ultimately be able to generate more pervasive stories. There are thousands of artists in towns continuing to generate them.

And soon, companionship, emotion, community, and sex!

About a dozen years ago I remember getting in this debate with someone about whether wearables would surpass iPhones (so far I think I won the debate) for a long time I think there’s been this vision of getting away from the computer, and at the time I said. People like TV is much to get rid of their iPhone. (https://www.derekthompson.org/p/why-everything-became-televi...)

I feel like for a long time people have been talking about trying to get rid of the computer and they haven’t yet. That technology wouldn’t be mediated by a screen.

The muddled point I’m trying to make is - I wish I’d lost the debate. I wish technology companies were focused on invisibly making things a little cheaper. A little faster. A little more convenient.

Revolutionize energy creation and efficiency, materials sciences

When I was a kid it was popular to say you shouldn’t watch too much TV it rots your brain. How could The people finding ways to eke every last bit of their tv into your worlds not be the villains?

I dunno, sometimes I wonder wha the point of all this progress is. what’s the point? What are we racing towards? At the end of all of this is death.

There are other philosophies and ways of living in the world and it feels like we’re abandoning them all for progress accelerationism. Even if that progress leads to our extinction

And I get a lot of use out of AI! But philosophically I would give it up in a heartbeat.

We don’t actually seem to have a coherent philosophy of what we’re doing here. So build AI to automate all the things, and then what? What happened to our ability to dream and make manifest a better life for ourselves?

Hope you’re not meeting with any mobsters. I think it’s not the aesthetic that worries me, it’s the functionality. Recording others should carry some friction and ceremony like pulling out a tape recorder. But maybe it’s just not the product for me.

Is there a light that tells people they’re being recorded? or is this like creeping-as-a-service.

Maybe it’s because the underlying economics have changed. The companies building these products are profitable and don’t need to rely on debt anymore.

https://www.reuters.com/business/retail-consumer/anthropic-t...

I might challenge the idea this is “smart politics”. Like I can get the tactics, but it feels disconnected from a greater strategy.

The best I can figure out is I guess good for the next two years having people subordinate to your administration.