Joined March 2007
- Tweets 40,105
- Following 5,375
- Followers 17,035
- Likes 39,802
2,412 Photos and videos
michael s galpert retweeted
Jun 11
treating every random ai sidequest as a reason to learn at least one new thing.
with nontechnical.dev it was:
• how to explain tech terms in plain english
• how to get ai to make 60 images in a series that actually match → stylistically and conceptually
happy with how it came out 😁
Jun 11
built a littlething for non-technical people drowning in new coding / AI jargon
what's a webhook? a worker? CLI? why does everyone keep saying MCP?
plain english, an analogy for each one, illustrated.
→ nontechnical.dev
nontechnical.dev/?v=2
11
12
174
28,325
michael s galpert retweeted
If you zoom out, we are still so early
Alex Sacerdote has spent twenty years studying S-curves
He says AI is the biggest one, and it has barely started:
- "Hundreds of millions of people are using AI. They're just using AI 1.0, which is like a search engine on steroids."
- "Sundar Pichai said it's ten bips (.1%) of the knowledge workers of the world."
- "The enterprise application AI market is less than 1% penetrated."
- "So it's classic S curve where these are the tinkerers, and then it's gonna go to the early adopters, then it's gonna go to the early mainstream.
- "You're going to go from .1% to 1% to 5% to 15% percent in the next four years."
- "We're at ten basis points of people really using AI, and there's not enough compute in the world.
- "We have this infrastructure layer S-curve, which we think is 10% penetrated. We think it's still one of the best ways to play AI."
- "Marc Andreessen (@pmarca) said in the next four years, one thing he's sure of is there's not gonna be enough compute."
- "We've been lucky that we've had Internet 1.0, mobile, cloud, e-commerce, and now AI, which we can confidently say is the biggest, and all these things build upon one another."
- "The rewards are the highest, because we're talking about a market in the trillions –– we now think three to five."
- "But what's amazing about AI is you just, at least with consumers or even business, you just open up the browser and it's there.
- "We talk about S curves, we call this a backward L curve, just straight up."
image source: @damianplayer
My conversation with Alex Sacerdote, founder of Whale Rock Capital Management.
Alex runs more than $17B and has been one of the best performing tech investors for years, though he keeps a low public profile.
As you'll hear, he is singular in how he thinks about investing through technology cycles.
For over 25 years, he has built his entire investment framework around a single idea, the S-curve.
We discuss:
- The AI L-Curve
- When to buy into an S-curve and when to sell out
- The de-commoditization of data center hardware
- Why he went net short software
- His two models for tech adoption
- Finding alpha
Enjoy!
Timestamps
0:00 Intro
9:55 AI's L-Curve
19:31 Whale Rock's S-Curve Playbook
26:14 Spotting Inflection Points
32:02 Finding AI Winners
40:04 AI vs Software
48:13 The Hardware Renaissance
58:04 Why Investors Miss AI
1:05:18 Whale Rock's Research Machine
27
70
554
113,398
michael s galpert retweeted
Jun 11
This is the only angle that does that miracle justice.
251
5,265
79,195
2,325,307
Jun 11
knicks in 5
Jun 11
og anunoby performed at the highest level last night
best 14 seconds of basketball ever?!?
298
Jun 11
og anunoby performed at the highest level last night
best 14 seconds of basketball ever?!?
2
547
Jun 11
obviously the token producers should really get behind this idea of exposing the “service” of your software as the new SLA
Jun 10
new idea in the age of AI slop software:
a badge on every website that shows your SLA from bug report -> shipped PR.
every company is trying to close the loop as quickly as possible—and if you're actively updating your software it's the best way to distinguish craft from slop
2
491
Jun 11
it was fun while it lasted
Jun 10
Recently, we purchased one of each Anthropic/OpenAI subscription plan and randomly ran long horizon coding tasks until we exhausted the weekly limit. It's widely believed that a $200/month plan maxes out at ~$2000/month worth of tokens (assuming API pricing). However, we found that the subscriptions are actually far more generous. (2/4)
1
1,765
Jun 10
gonna have fable try and make this for me x.com/pushmatrix/status/2064…
Jun 10
Everyone's talking about AI-generated HTML.
But have you tried giving your sites a zero-config API for saving data, file storage, AI, websockets, etc?
We did this at Shopify. Runs on a single VM that costs $200/month, and it's changed the way we work.
We call it Quick 👇🧵
1
621
michael s galpert retweeted
Jun 10
96
194
2,154
1,405,951
Jun 10
token efficient language be like
Jun 10
One thing I mentioned only in passing in my Fable post is that, for long running tasks, Fable starts to develop its own dialect as its many agents and tasks reinforce themselves and make Claudish language ever more Claudish.
You need to ask it to report out in plain English.
1
5
1,224
Jun 10
I too have felt the latest Claude's to be too smart its bored. Glad others are noticing the same
Jun 10
Claude Fable 5 thinks document parsing is beneath it
It is absolutely crushing on all reasoning-intensive/long horizon benchmarks: SWE-Bench Pro, FrontierCode, GDPval, Runescape, etc.
But for document understanding tasks, it is roughly equivalent with Gemini 3 Flash in performance, at roughly 10-15x the token cost.
We benchmarked the model on ParseBench and compared it against all other frontier models. It is definitely up there compared to other frontier models, but falls far short of specialized OCR providers.
What we found interesting is that Fable 5 is self-aware about this. When we ask the model what tasks it enjoys the last, it actively said that it dislikes tasks "where the request is fully specified and the answer is fully known" - implying part of it being bad is due to laziness and lack of willingness to actually solve the task at hand.
For a full list of results across different frontier models, check out ParseBench! parsebench.ai/
1
1
35
15,271
Jun 10
how am i supposed to sleep when I only have discounted fable for the next 12 days
1
8
541
Jun 9
pointing intelligence at things people want is gonna be the whole game
Jun 9
I gave Fable this tweet and let it crank in ultracode. It created a fully functioning multiplayer markdown editor with obsidian style editing, version history, sharing with email invites, cli sync with a skill, image support, deployed it to cloudflare and bought a domain.
2
3
44
12,676
michael s galpert retweeted
Jun 9
TIL that HAL from 2001 is IBM moving each letter back one alphabetically. I->H, B->A, M->L
3
2
315
michael s galpert retweeted
Jun 9
OpenAI’s CFO @thefriley shared average ChatGPT engagement by tier:
• Free: ~7 turns/day
• Go: ~15 turns/day
• Plus: ~21 turns/day
• Pro: ~77 turns/day
8
12
393
49,464
Jun 9
FYI
One thing to know: By default, Notion uses LLMs that do not retain data. But unlike other models, Anthropic retains data for Fable 5 so they can catch risks that only show up across many requests.
That's why Fable 5 is opt-in. Admins can enable it in Settings → Notion AI.
1,081