Ethan Mollick Profile picture
Mar 3, 2022 8 tweets 6 min read Read on X
My sister @JMollick produced #TheDropout - the new Hulu series on Elizabeth Holmes. In addition to being entertaining, it also shows some drivers of success in entrepreneurship.

So: a 🧵 of research on Theranos, and what honest investors & founders can learn from the lies. 1/
Much of the fraud was explained by "Symbolic Action." In a classic paper, Zott & Huy find that founders who skillfully use symbols do better, since folks view the symbols as indicators of real ability. They identify four categories of symbolic action, all exploited by Holmes 2/
The first category is showing personal capability, and the paper describes multiple ways of doing this: you can look the part of the entrepreneur; you can conspicuously show connections to top schools; or you can show that you are personally "all in." Elizabeth did all three. 3/
The 2nd category is showing that you are organized like a real professional organization. Common ways for entrepreneurs to indicate this are to have professional office spaces and the trappings of what people expect to see in a real firm. Elizabeth was very aware of this. 4/
The third category is to show symbols that your business can achieve its goals. The three classic ways to do this are to show off half-working prototypes, win industry awards, and show that you have received money from prestigious funders. 5/
Finally, we have a demonstration of key stakeholders, because if important people back your company, it must be good right? See the Theranos Board! (Of course, this didn’t convince real biotech VCs, who would have wanted to see stakeholders in the medical field.) 6/
And Holmes also was very good at pitching. This paper shows how she pitched Theranos using powerful techniques:
🖼Framing: Why the world needs improvement
💉Filling: Vivid images of how she would solve it
👥Connecting: Showing others trusted her
💪Committing: Showing dedication
7
Unethical startups are more likely to raise 💰 but also tend to waste it, hurting overall innovation. By comparing 2 sets of books, this paper identifies Chinese startups that got grants via fraud. Frauds were less likely to hire quality people & to conduct significant innovation

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Ethan Mollick

Ethan Mollick Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @emollick

Dec 17
Veo 2 prompt: "a distant shot zooms in to reveal a knight wearing a golden helmet, he begins to charge on his zebra, lowering his lance, charging towards a clockwork octopus" (this is one of the initial 4 videos it made)
"an woman with short black hair assembles an impossibly complicated device, close up on her face, she is sweating"

The consistency of small details is really impressive, the fact that the shaft of the screw turns at the same speed and direction, hair and sweat, tattoos...
"a man holding the leash of a golden retriever stares mournfully at a fireworks display over his small town of Tudor-style homes, the flashes punctuate the darkness."

All videos are from the first 4 from the prompt. I did learn that you can't ask for many cuts or scene changes.
Read 10 tweets
Dec 8
I thought puzzle design beyond current AI, so this is the first time I have seen a model actually create a solvable and interesting puzzle.

"o1, create a d&d puzzle that is non-trivial, but solvable. it should not involve elements or statues, be creative."

(Solution in thread) Image
Here are the hints it gave and the solution.

(There is one unnecessary step in the puzzle, by the way, but a zero shot coherent word puzzle is pretty neat) Image
Image
I suppose the silent e's might complicate things, but I wouldn't think they would stop solving.
Read 5 tweets
Oct 26
Claude: “Give me hard original writing prompts for an MFA program” yields some really clever (and near impossible) prompts. Image
Image
Image
A few more. (Literal prompt: “give me some more”) Image
Image
Image
Image
It did well on #12 (broken into parts)

# The Night Shift
## Operation Manual: TX-2047 Industrial Processing Unit
### Document Classification: Essential Reading
### Last Updated: 27 October 2024

**WARNING: Failure to follow proper shutdown procedures may result in catastrophic system failure and/or permanent spiritual contamination of the processing unit. All operators must complete this training module before attempting solo operation.**

I've been working the night shift at the plant for three weeks now, and I still can't shake the feeling that something's wrong. The constant hum of machinery should be comforting – white noise to fade into the background – but instead it feels like a presence. Watching. Waiting.

**1. INITIAL POWER-UP SEQUENCE**
- Locate the main control panel in the eastern wing
- Insert operator key and turn clockwise until resistance is felt
- Wait for the safety lights to pulse three times

The eastern wing is the worst. During day shift, it's just another industrial corridor: concrete floors, steel pipes, fluorescent lights. But at night, those same lights flicker and dim, casting shadows that seem to move when you're not looking directly at them. The shadows are always darker near the main control panel.

**WARNING: Never attempt to force the operator key if resistance occurs before a full clockwise turn. This may indicate a structural compromise in the primary containment system.**

The manual says the resistance is normal, just the safety systems engaging. But it feels different every time – sometimes a clean click, sometimes a grinding that sets my teeth on edge. Last night, I swear I heard whispering when I turned the key.

**2. PRIMARY SYSTEM ACTIVATION**
- Check all pressure gauges in sequence (A through F)
- Verify each gauge reads within acceptable parameters (see Table 1.1)
- If any gauge shows readings in the red zone, proceed to Emergency Protocol 7

They never told me what happened to the last night shift operator. Clean record, five years of experience, then one morning they found the control room empty. The key was still in the panel, turned halfway. All six pressure gauges were pegged in the red.

**Table 1.1 - Acceptable Operating Parameters**
```
Gauge A: 2.3 - 2.7 kPa
Gauge B: 15.0 - 15.5 kPa
Gauge C: 7.2 - 7.8 kPa
Gauge D: 3.0 - 3.3 kPa
Gauge E: 9.1 - 9.4 kPa
Gauge F: 4.5 - 4.8 kPa
```

The numbers dance in front of my eyes during every check. Sometimes they seem to change when I blink, but when I look again, they're normal. Always just within acceptable parameters. Always just barely safe.

**3. COOLANT SYSTEM INITIALIZATION**
- Open primary coolant valves in the following sequence ONLY:
1. Valve 2C
2. Valve 4A
3. Valve 1B
4. Valve 3D
- Listen for the distinctive activation tone
- Verify flow rates on digital display

The activation tone. That's what they call it in the manual. A clear, mechanical chime that means everything's working properly. But it doesn't sound like a chime anymore. It sounds like singing. Like a voice calling from somewhere deep in the pipes.

**WARNING: The correct valve sequence must be followed precisely. Improper activation may result in system contamination and anomalous behavior of both mechanical and digital components.**

I asked my supervisor about system contamination once. His face went pale, and he told me to stick to the manual. Just follow the procedures. Don't ask questions. Don't investigate unusual sounds. Don't look too long at the shadows near the pressure gauges.
Read 5 tweets
Oct 21
I wish people would stop repeating these as if they are facts that AI is plateauing.

AI might hit a roadblock, we don’t know, but every one of these issues has multiple studies stating the opposite: synthetic data works, scaling is fine, etc. We need more nuance on the AI future Image
Image
Image
One paper arguing the other side of the “LLMs don’t reason” debate
Read 5 tweets
Sep 27
It is amazing how the industry came together to invent a universal USB-C connector and then decided to instead make it all a giant mess. Image
Remember to research your cable as well. Image
At least the symbols are intuitive. So easy to figure out which carry power and which carry data.

Amazing choices. Image
Read 5 tweets
Aug 18
I don’t know if LLMs can reason in theory, but they seem to “reason” in practice.

Ask Claude a Fermi problem: How many telephone poles are there in Manhattan? (Guess: 440. Actual: 0). On Staten Island? (Guess 19000-2859. Actual: 27,137) Queens? (Guess: 51-85k. Actual: 103k)

Image
Image
Image
It doesn’t seem to have the number handy, but, even if it did, next token prediction should mean it reasons forward, not backward from a conclusion.
They are obviously not good at all sorts of problems (logic puzzles, etc.) but they do well at many others (Fermi problems, etc).

Reasoning is a jagged frontier for LLMs.
Read 6 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(