Speaking of which, when did the boards get the little “summarize with AI” button. Does it cost @randy 5 bucks every time we press it?
It’s probably a “free” add on that is nigh-impossible to shut off, just like the ones Google and Microsoft are shoehorning into everything.
The earth still pays as gallons of water are irradiated to summarize my thoughts on Elden Ring to disinterested parties. 
3 Likes
5 Likes
I’m just glad that as long as Gmail is going to try to force its way into my compositions, its little pop-up option now says “Help me write,” as opposed to it saying “Polish” before.
I legit couldn’t understand why it wanted me to write in the Polish language.
7 Likes
2 Likes
It’s time to invent some suitably relevant company names:
fAIk
fAIq
flAIke
dAIlight robbery
and for the accounts payable departments:
unpAId and unashAImed
too big to fAIl
your cheque is in the mAIl
<user name> is on vAIcAItion
somewhere in that AI is some artificial intelligence
and the supply chain of a trillion dollar company
2 Likes
Please be pAItient while we direct your cAIl to a customer servAIce representAItive
We apolAIgize, your clAIm has been denAId
This is a worthwhile read to understand the AI hype and the fallacies it contains:
The whole premise is that AIs are rapidly deflating the value of knowledge and expertise on any given matter.
While AIs are not comparably smart in bits-per-byte view with humans, they do close the gap in some cognitive tasks like knowledge retrieval by having several orders of magnitude more communication bandwidth. Frontier AIs like ChatGPT are knowledge compressions of the Internet.
Simplifying PhDs to textbook experts on a task and nothing more is unbelievably stupid. PhDs also expand the knowledge on a given field, they help humanity discover new areas of knowledge in that field.
This entire discussion about AIs matching the intelligence of the ‘best of the best’ humanity can offer also clarifies how, once again, we are framing intelligence in an entirely wrong fashion.
Models only perform well in situations they’ve seen or experienced beforehand. In other words, they depend highly on their capacity to fetch the reasoning pattern from memory instead of ‘acting it out.’
In reality, most of their performance is due to memorization and not actual reasoning.
Therefore, seeing Deep Research tools and claiming that this kills knowledge work, PhDs, or whatever you are interested in killing today to gain attention and clicks, is simply not true.
While there are some reasoning models in the early stages, it’s not clear if that ‘reasoning’ is truly closing the gap, or only a more advanced way of knowledge retrieval. In other words, can these models actually act out new thoughts?
1 Like
Anyone tried this yet?
Not something I will touch out of principle!
3 Likes
Here’s a fun experiment: https://www.youtube.com/watch?v=160F8F8mXlo
Good demonstration of the inherent limitations of genAI.
Also, we’ve been having an interesting chat about the topic of AI in film and advertising over on CML with the DPs: What’s so AI about this?
Just read this, and will check out the podcast. I’m conflicted on Nadella’s true intention: Is he trying to lower expectations or is he trying to be a voice of reason in a sea of tech-bro hyper-valuations? There’s some logic to the pov that investors might see MSFT as a safer investment if the CEO is making more realistic claims.
I think Satya is saying, no “revenue” value, but not realworld value. The infrastructure to train and infer is very expensive, and the monthly subs have not yet become profitable.
1 Like
Not gonna use the AI by and for nazis. I’ll give up on computers entirely before that.
7 Likes
Based on my hackneyed Super-Bowl-Ad-Based reasoning he is trying to set up an offramp. Two years ago the Super Bowl was awash in crypto ads. Reporting on the matter said it was because people wanted to get their profits out so they needed new people to buy in.
This year we had gobs of AI ads. Big Tech needs this lightning to strike, and those ads are a Hail Mary pass. The only mention of AI I hear in normal people discourse is, “how do I turn this shit off in Microsoft/Google products?”
As a side note, I love how AI ads present protagonists that are fully incompetent. Not only do they not know shit, they can’t even find info on the internet. They need a special internet nanny to tell them how to change a tire. I’d pay good money to watch two dozens of these folks in a season of Survivor.
OH SHIT: Million dollar idea: survivor style reality show where the AIs everyone use on the island are just real people.
Anyway, yeah, this dude isn’t dumb1, this dude sees the real numbers, not the PR numbers.
1. Not a given that a CEO is smart but I’m giving Captain Microsoft the benefit of the doubt until he posts a two-shield carrying, fat rolling, talisman cancelling Elden Ring build.
Nice, they can put the servers in all the We Work spaces that SoftBank dumped cash into.
It wouldn’t be that different to the number of “startups” that put servers and storage into the bathrooms of apartments that they rent for business purposes…
i’m sure that you remember one or two of those.
in 1999 one of the tezros that i was working on caught fire in what was laughably called a machine room.
god bless imaginary fArces…
2 Likes
We 100% had a rack of flame and decks in a bathroom. We also realized that we could remove the drainpipe from the tub, run cables down it and out the little service panel into the room below.
I genuinely don’t know if anyone ever thought to shut the water off to the tub, but it never ended up being a problem?
1 Like
Yeah, I agree. I did like his line about when AI means 10% sustained productivity growth we’ll know it’s successful, or something like that.