“The personal computer, like television, can greatly increase the forces of both good and evil.”
— Nils Nilsson, SRI, 1982
Jerry Neumann’s Tweets
The iPhone is not descended from the Apple II, it is a descended from the HP-65
1
4
5/ Apple and Microsoft IPOed in the early 80s bubble… as did several hundred tech companies that subsequently went bust
This essay from is by far the best chronicle I’ve found of the period
1
7
79
Show this thread
Great article by my friend
2
1
I report, with great sadness, the passing of Paul David. A fabulous scholar of economic history and the economics of technology, he lit up Stanford for six decades. siepr.stanford.edu/people/paul-da
19
174
441
This is such a dumb way to look at this. Neither was the microprocessor particularly innovative when introduced by Intel. What matters is not how revolutionary the tech is, but how revolutionary the uses are
3
6
19
It's almost a rule that, when founders cold email me and say I should be interested because I invested in ___ and ___, they inevitably pick my worst-performing investments
4
2
32
I have 2 physical books, rather large, with a bunch of tables in them. What's the best way to get it into a digital excel type format? Happy to pay if I could figure out whom.
6
2
11
Use this one instead, it reads better
5
Show this thread
Aside from me switching from conciseness to concision. I'll fix that
Show this thread
Drafting ChatGPT/LLM guidance for my class (see below.) Thoughts?
8
2
22
Show this thread
The question here, really, is: what is valuable about predicting the future of technology?
7
3
12
Show this thread
It is 1970 and you have a way of predicting with relative certainty 2023's internet. What would you do about it?
NB: you could predict the general shape of things--global interconnection, digital everything, etc) but not the specific companies and people
22
7
28
Show this thread
Thanks for the shout-out and nice summary of the book's message!
Quote Tweet
“Why Managers Matter” authors Nicolai Foss and Peter Klein say hierarchical management structures are still “the worst form of organization—except for all the others.”
The challenge: constructing a well-functioning hierarchy.
Theodore Kinni reviews: sb.stratbz.to/TK-cbmw
2
8
Oh oh oh. I've got the new business model. You have advertisers pay you to put objective facts into your database. Like, Fact: Best soda=Coca Cola. You charge on a cost-per-fact basis. It's the perfect business model for our times!
4
13
Show this thread
Ancillary question: couldn't you train a LLM to output tokens for text mixed with tokens for a database search, to prevent hallucinations on objective facts?
3
3
Show this thread
(I guess I shouldn't anchor my question with my assumptions about the answer, but it really is a real question. I am prepared to be schooled, that's my job.)
2
7
Show this thread
This question is aside from all the technical difficulties of LLMs (hallucinations, time-to-update, etc.) I'm assuming these can be fixed over time (though I'm not so sure the can with the current tech, given how neural networks work)
1
4
Show this thread
I get that if I ask an LLM for a poem about feet in the style of Wordsworth, there isn't an opp for ads but, then, there never was, was there?
1
7
Show this thread
So, if I search for 'best mortgage rates in NJ' and I get an answer in text form from an LLM versus a bunch of three-line answers, won't I still click through, if they're relevant?
1
6
Show this thread
Couple this with *who* pays for ads: companies that want people to click-through and end up on their landing pages.
1
6
Show this thread
As background...the reason Google won over the incumbents 20 years ago was they rejected the portal approach of Yahoo! and Excite in favor of actually giving relevant, want-to-click-through results. Then charging for click-throughs.
1
6
Show this thread
I keep seeing this claim that Google is vulnerable to LLMs because long-form answers are not amenable to advertising. I know something about advertising, but I can't get this claim to square with what I know...can anyone with actual (cpc) advertising experience set me straight?
11
2
42
Show this thread
Happy new year!
2
13
Show this thread
2023 will be the year that my book will be published after more than five years of (hard) work.
(please retweet)
25
162
803
Interesting that, for the hardest problems, AI does better with a massive training set than deducing from assumptions, while humans think they should spend their time reasoning from first principles rather than amassing a bigger training set
3
28
Last breakfast club of the year - such a good morning
Big thank you to my cohosts and
4
5
38
Show this thread
I am super duper excited today to announce Coelius Capital 2. A new $33.3m fund for investing in early stage B2B software. Alongside Coelius Capital Opportunites 2,my rolling fund and the Angellist syndicate it comes to about $100m in new capital to put to work.
131
44
874
Show this thread
1/ EOY prediction time! 2023 will be the year of racing to incorporate AI in products, but most companies' first instinct (shallow integration of generic models available through limitied APIs) is only table stakes.
The race is now on to:
27
98
624
Show this thread
As a contrarian, I am now looking for investment opportunities in honey
3
14
Jeez, it's like this thing is reading my mind
2
2
11
But comforting that even it hasn't finished The Pale King (though its excuses remind me of some of my students' excuses)
2
4
Show this thread
It's sort of embarrassing that GPT3 is better read than I am
2
6
Show this thread
Quote Tweet
Quitting has a bad rap, but it can often be a smart, strategic move.
Nobody knows that better than @AnnieDuke, a former professional poker player. She joined @krysboydthink to explain how to know when to fold 'em. Listen here! think.kera.org/2022/12/01/a-p
2
4
12
The joy of editing.
2
13
My friend Liz Zalman has written what looks to be a great book with on the founder/investor situation:
smile.amazon.com/Founder-Vs-Inv
4
6
5