Eugene Mandel

Ex-Head of AI at Previously Head of Product @expectgreatdata. Product leadership, startups, machines + humans.

[Alex Sherstinsky @AlexSherstinsky]

Conversation Analysis reveals what truly unfolds in business conversations — far beyond the binary "won/lost" label common in sales. A powerful reminder that the "best" outcomes often transcend immediate closes, turning apparent losses into stronger, trust-based wins later.

[Eugene Mandel @eugmandel]

@carlvellotti CC

[Carl Vellotti 🥞 @carlvellotti]

Every PM should be using Claude Code.

So I built a HUGE course for you to learn Claude Code... IN Claude Code!
🔹 Complete guide
🔹 Make PRDs, analyze data, create decks

Soon, I'll sell it for $149.
For the next 24h: FREE!

Follow + RT + comment "CC" & I'll DM it.

[Eugene Mandel @eugmandel]

@Zephyr_hg NEWS

[Eugene Mandel @eugmandel]

Today, this paper from almost 30 years ago has clear practical value, as doing this right might be the difference between success or failure of voice agents that do customer service, sales, and other conversational tasks.

[Alex Sherstinsky @AlexSherstinsky]

7/7 Read more about different ways of materializing speech acts to imbue AI agents with personality in this 1997 (correct! - AI agents are not a 2024+ invention!) paper by Janet E. Cahn et al (https://t.co/JoNyB1224a). #ConversationalAI #ConversationDesign #ConversationAnalysis

[Eugene Mandel @eugmandel]

Understanding that conversation analysis of voice calls can achieve good enough quality by relying on transcripts alone saved us a lot of time at Loris when we were expanding from supporting digital (chat like) channels to voice. We could reuse all the technology we already had.

[Alex Sherstinsky @AlexSherstinsky]

You've probably heard:
"93% of communication is body language." It’s repeated in keynotes & sales trainings. But it’s misleading. Here’s its origin, why it stuck, and what really matters in human communication.
#ConversationalAI #ConversationDesign #ConversationAnalysis

[Eugene Mandel @eugmandel]

At the first glance, this is such a weird feature of human conversations. But it is not that weird. The speaker is managing both the attention of the other participant, making sure that they are ready to receive their own and their own fear of rejection by a kind of warm up.

[Alex Sherstinsky @AlexSherstinsky]

1/2 "You know what I’m thinking?" isn’t a telepathy quiz — it’s a pre-action: a conversational move that sets up what’s coming next.

[Eugene Mandel @eugmandel]

And of course when the participants ignore the real meaning, it creates funny moments: Q: "Can you hand me the pen?", A: "Yes, I can" (no action).

[Alex Sherstinsky @AlexSherstinsky]

2/3 If someone says, "Can you hand me the pen?", it is not a question about ability, but a request. #ConversationalAI #ConversationDesign #ConversationAnalysis

[Alex Sherstinsky @AlexSherstinsky]

Word choice matters: Saying "I want to talk to you" in a tense and strained situation may be rejected as intrusive. However, saying "I want to speak with you" would be perceived as respectful and thus accepted. #ConversationalAI #ConversationDesign #ConversationAnalysis

[Eugene Mandel @eugmandel]

my deep thinking over a box of documents for shredding

[Alex Sherstinsky @AlexSherstinsky]

Experts live in one conversational world, lay people in another: https://t.co/IRuf2srX6E #ConversationalAI #ConversationDesign #ConversationAnalysis

[Alex Sherstinsky @AlexSherstinsky]

The "use their name often" sales tip is a myth. Fake familiarity from reps feels inauthentic—we don't know each other. Skip forced rapport, just help accomplish the goal. #ConversationalAI #ConversationDesign #ConversationAnalysis

[Alex Sherstinsky @AlexSherstinsky]

To converse is human. Thoughtful, intentional conversations can save the world. Reckless conversations antagonize and cause conflicts. The book "Talk" by @lizstokoe elucidates just how much we act through words. #ConversationalAI #ConversationDesign #ConversationAnalysis

[Eugene Mandel @eugmandel]

@huang_song_ @Jaxtr

[Huang Song @huang_song_]

Get a https://t.co/1Z3AGpyTeg invitation code now!

Just:
1. Like this post
2. Repost
3. Tag a friend below!
4. Follow me so I can DM you the invitation code.

[Predibase by Rubrik @predibase]

What if...
➡️ you could train an LLM to beat GPT-4 for your use case?
➡️ and, that LLM was open-source and <1% the size of GPT-4?
➡️ and, you could do it for free?

If that sounds appealing, then don't miss our virtual workshop - free compute included!

https://t.co/xSONeCWSZg

[Eugene Mandel @eugmandel]

@brianwilt @pitdesi Wow - our former office is involved in a criminal plot! :)

[Ethan Mollick @emollick]

A clear use case for AI is idea generation. The evidence is strong across many studies that AI (GPT-4, Claude 3, Gemini Advanced) can help boost human creativity & the costs of getting bad answers is low.

My Innovator GPT: https://t.co/Jha0oT9DIj

More: https://t.co/bXfPZaRl5f

[Sumit @_reachsumit]

Is Cosine-Similarity of Embeddings Really About Similarity?

Netflix cautions against blindly using cosine similarity as a measure of semantic similarity between learned embeddings, as it can yield arbitrary and meaningless results.

📝https://t.co/rbtsmXQ19s

[The New York Times @nytimes]

New details show a pattern of rape and mutilation against Israeli women by Hamas in the Oct. 7 terrorist attacks, a New York Times investigation found. https://t.co/FrChcRLvYC

[Russ Roberts @EconTalker]

Always thrilled to learn something new from an EconTalk convo. This week's episode with @havivrettiggur is special because I learned many things. I gave it an unusual title because I don't want you to miss it. Please RT: https://t.co/TM8MeK8E5Y