Pedro Tabacof

91 posts

Pedro Tabacof

Pedro Tabacof

@PedroTabacof

Staff machine learning scientist at Intercom.

Dublin City, Ireland Katılım Eylül 2012
982 Takip Edilen168 Takipçiler
Pedro Tabacof retweetledi
Mohit Hajarnis
Mohit Hajarnis@HajarnisMohit·
One of the often slept-upon benefits of attending the University of Chicago is that they make you read Marx as part of the core curriculum, which is why this article gave me flashbacks of taking SOSC 114 as a freshman. Marx, writing during the Industrial Revolution, predicted capitalism would periodically devour itself: firms replace labor with machinery to boost profits, but competition diffuses the technology, drives prices to marginal cost, and the gains get competed away. Meanwhile, displaced workers lose purchasing power, hollowing out the demand the whole system depends on. Production rises but no one can afford to buy what's produced - the contradiction between production and realization. Citrini's piece describes this exact dynamic, then declares there's "no natural brake." It's the most Marxist piece of financial analysis written in years, and makes the same errors Marx did. Schumpeter offered the obvious rebuttal 80 years ago: creative destruction doesn't just destroy, it creates industries we can't yet conceive of. Everyone in the replies is already making this point, and I think they're right. But the sharper rebuttal is Hayek's: prices are the brake Citrini says doesn't exist. Who funds $200bn / qtr in AI capex when equities are down 38%, private credit marks are in the 50s, and consumer demand has collapsed? Cost of capital rises. Incremental build-out becomes uneconomical. Capital gets destroyed and reallocated. Citrini also unknowingly describes Marx's proletarianization of the petite bourgeoisie: the $180k PM driving Uber is textbook. But the article claims this collapses consumer demand, and that's where it breaks. The top decile drives 50%+ of spending and their wealth is in equities, not W-2 income; they're long the hyperscalers posting records in Citrini's own model. Blue collar is insulated because AI replaces cognitive labor, not physical. The professional middle class gets crushed, but aggregate demand doesn't. The spending class IS the capital-owning class. The K-shaped recovery they fear actually stabilizes the demand base they say is collapsing. In the stable aggregate demand, the petit bourgeoisie finds ways to reinvent itself. I think the Citrini piece is excellent and worth reading. But history has repeatedly shown that periods of transformative productivity gains ultimately accrue to the consumer through lower prices, more leisure, and higher quality of life. Marx's error wasn't diagnosing the disruption, it was underestimating the system's ability to adapt.
Citrini@citrini

I spent 100 hours over the past week researching, writing and editing the piece we just put out. It’s a scenario, not a prediction like most of our work. But it was rigorously constructed, dismissing it outright requires the kind of intellectual laziness that tends to get expensive. And we’ve released it for free. Hopefully you enjoy it. citriniresearch.com/p/2028gic

English
194
610
6.2K
1M
Pedro Tabacof retweetledi
Intercom
Intercom@intercom·
Some questions are just easier to answer visually. @Fin_ai can now send screenshots and GIFs from your existing support content, giving customers step-by-step visual guidance exactly when they need it. This works automatically across channels, with no setup required. @PedroTabacof shows how it works in practice, and why it’s such a powerful upgrade for customer clarity.
English
1
2
7
623
Pedro Tabacof retweetledi
Pratik Bothra
Pratik Bothra@pratik60·
@Fin_ai (Intercom’s AI Customer Agent) has officially crossed 10 trillion tokens processed in production. This excludes backtesting, development traffic, and custom model usage. In @intercom, we split workloads across multiple vendors to optimize for cost, performance and reliability. That’s millions of conversations, decisions, and instructions interpreted in natural language.
Pratik Bothra tweet media
Deedy@deedydas

OpenAI just dropped the top 30 customers who've used 1T+ of their tokens.

English
1
7
32
8.5K
André Castro
André Castro@andrefrcastro·
@PedroTabacof @destraynor @Fin_ai @eoghan Oh man, it works so well. This custom context lets fin dig in some deep problems that would require a human to check on them. You should automate/template custom params on the guidance UI. Many customers don’t know what they are missing
English
1
0
4
68
Des Traynor
Des Traynor@destraynor·
We're putting our money ($1M) where our mouths (or more specifically our resolution rates) are with @Fin_ai - read @eoghan's post for the details.
Eoghan McCabe@eoghan

Today we’re launching the Fin Million Dollar Guarantee. We’ve been making big claims about our industry-best performance, but we know such claims are cheap. So it’s time to put our money where our mouths are. If we don’t deliver on these claims, we’ll pay you $1M. First, some background: In the very near future, all of your customer operations will run on AI. All customer touch points, at every stage in the customer lifecycle, from marketing, sales, service, and success, will be owned by a perfect, charming, expert customer agent, engaging with every customer and available instantly every hour of the day. You’ll choose one vendor for all of this, not many. And you’ll choose the vendor who can do the most high quality work. For the vendors, the opportunity is massive. We’re talking about disrupting literally trillions of dollars of human work, and potentially chasing TAMs 25-50X the size in SaaS. So, unsurprisingly, this space is extremely hot, well funded, and fiercely competitive. For this reason, I’m about to be very honest with my competitive analysis and say the things out loud which every confident CEO says behind closed doors. But please don’t imagine I’m attacking anyone. I know a lot of the CEOs in this space, and admire all of the companies pushing this category forward. The new guys in particular are doing a very impressive job, and it’s generally assumed that they’ll eat the lunch of older incumbents like us. At the risk of over simplification, and to cut to the chase, I believe there are only three players in this space who matter: our product @Fin_ai, and newcomers @SierraPlatform, and @DecagonAI. Sierra is taking a consulting oriented approach with non-tech enterprise customers. This will be extremely lucrative in the near term, but I’m bearish on this approach in the longer term. In the early days of the web, large consulting companies made the exact same enterprise play, charging millions of dollars to deploy and run web sites. But web sites soon became an integral part of business operations and all enterprises eventually wanted to manage them in-house. The same will happen with agents, and in-house teams will need a management application layer, which we’ve made extensive investments in. With Fin we’ve started slightly lower in the enterprise layer, with tech or tech-forward companies who are ready for more sophisticated deployments of AI, which they want to manage themselves. We’re competing directly with Decagon for this market. And while the customers we’re winning against them cite our uniquely deep management application layer, it’s our superior performance that delivers the most striking and immediately valuable difference. It’s become very clear that our AI Group has built highly differentiated tech, which we’ll reveal more about this summer. The result is that we’ve beaten Decagon on performance in 100% of bake-offs that we’ve engaged in and have been chosen as vendor of choice 100% of those times too. Take a representative case of [very large SaaS unicorn], who was an early reference customer of Decagon’s. In a customer-run bake-off, Fin achieved a resolution rate of 63% (which has continued to increase to 72%, as of last week), while Decagon achieved a resolution rate of 49%. That’s 23 percentage points of additional customer requests which now receive instant, perfect resolutions to their problems. And 23 percentage points of additional customer requests that they pay just $0.99 per resolution for, instead of the $15-20 per resolution they’d have to pay humans for. With a performance gap this wide, your choice as a customer is made for you. This dynamic against them, and similar players in the market, has made Fin the largest by customer count (>5k) and largest by ARR (8 figures, which will cross $100M <3 quarters from today) in the category. ARR from our existing customers is growing at 4X, but incredibly, from new business customers it’s growing at 5X. We’re also rated #1 in the category on G2 and shipping every major new innovation in this category before everyone else. So, to the guarantee. Starting today, we promise two things: 1. If you sign up for Fin and are not 100% satisfied in your first 90 days, we will give you up to $1M of your money back, no questions asked. 2. If you sign up for our Fin Guarantee Success Program, and we do not achieve at least a resolution rate of 65% (which our studies have shown is the resolution rate of humans), we will pay you $1M! To learn more about the guarantee and to apply for the program, please visit: fin DOT ai SLASH guarantee :)

English
6
6
58
8.8K
Pedro Tabacof
Pedro Tabacof@PedroTabacof·
@andrefrcastro @destraynor @Fin_ai @eoghan Fin will take those attributes into account when crafting its answers. How much that will improve them needs some testing. Tbh, I haven't seen many applications like yours, but keen to hear about the results.
English
1
0
1
89
André Castro
André Castro@andrefrcastro·
@PedroTabacof @destraynor @Fin_ai @eoghan Thanks, really helpful advice. You should add it to the guidance UI. does Fin actually use rendered user/company attributes to improve response quality when the guidance refers to them? Or is it not really worth including that kind of contextual info in guidance?
English
1
0
0
96
Pedro Tabacof
Pedro Tabacof@PedroTabacof·
@andrefrcastro @destraynor @Fin_ai @eoghan Hi André, I work with the AI side of Fin. Here are two tips: 1. Long Guidance with loads of attributes might be hard for Fin to follow, so I'd recommend breaking it down into smaller pieces. 2. LLMs are bad at date maths so they might not quite get what "date is near" means
English
2
0
2
119
Pedro Tabacof retweetledi
Paul Adams
Paul Adams@Padday·
New Fin AI blog! A few weeks ago we decided to build a lot more in public. With AI, we're at the equivalent of 1999 for the internet, and most of the invention is yet to come. All of us have a lot to learn, and a lot to gain, by sharing what we know with each other. Today we're launching a new AI research blog, written by our excellent AI team. It is really good work. We're learning so much building Fin, so please share it with your teams building AI products. I bet they will find it valuable! fin.ai/research/
English
4
11
43
3.6K
Pedro Tabacof retweetledi
Fabrício Carraro
Fabrício Carraro@fabriciocarraro·
No episódio de entrevistas dessa semana do podcast IA Sob Controle, tivemos como co-host um amigo do podcast, nosso caro Ahirton Lopes (@ahirtonlopes), e nós 3 conversamos com o Pedro Tabacof (@PedroTabacof), Staff Machine Learning Scientist na Intercom em Dublin, na Irlanda - uma empresa que vem adotando e apresentando soluções envolvendo IA e agentes em tarefas de atendimento e relacionamento com o cliente. open.spotify.com/episode/4J8aoW…
Português
1
2
6
501
Pedro Tabacof retweetledi
Ciaran Lee
Ciaran Lee@ciaran_lee·
I spent 5 mins setting up Fin for @remarkablepaper, then asked it a few questions alongside their Agentforce setup. One of these agents is doing a far better job than the other.
Marc Benioff@Benioff

Agentforce is live at @remarkablepaper, where AI and humans now unite to revolutionize customer success. This is just the beginning—Customer 360 + Data + AI is redefining what’s possible. Excited to see Remarkable leading the way! ❤️🤖

English
5
8
79
17K
Pedro Tabacof retweetledi
Intercom
Intercom@intercom·
G2 just named us the #1 AI agent on the market – and it’s all thanks to real feedback from people who actually use Fin. No fanfare. No empty promises. Just proof that AI is making a difference where it counts, helping CS teams do more, faster, and better. And we’re not stopping here. We’ll keep building the best bots in the business so your support teams can keep winning.
Intercom tweet media
English
0
5
14
1.1K
Pedro Tabacof retweetledi
Intercom
Intercom@intercom·
"Working with @AnthropicAI keeps us at the forefront of AI, letting us explore the future and work on the bleeding edge, while delivering stable, reliable, and incredible results for our customers.” – @destraynor anthropic.com/customers/inte…
English
0
3
10
1.7K
Pedro Tabacof
Pedro Tabacof@PedroTabacof·
@mayurc137 @intercom @Padday @destraynor Note that all the information Fin used, including the name Hoban Washburne, comes from that article I linked above. We will see how we can improve such cases, as there was no need for Fin to search in the help center for such a simple question.
English
0
0
1
54
Pedro Tabacof
Pedro Tabacof@PedroTabacof·
@mayurc137 @intercom @Padday @destraynor Hi @mayurc137, I am from the Fin AI team, thanks for reporting it: I looked into it and, while a bad response from Fin, there is no personal informational being leaked. The information Fin used here comes from this developer article Fin has access to: #example-object" target="_blank" rel="nofollow noopener">developers.intercom.com/docs/reference…
English
3
0
1
121
Pedro Tabacof retweetledi
Eoghan McCabe
Eoghan McCabe@eoghan·
Every other month, we package up all the new tech we've just built and bring our product leaders together to demo it to our customers and the world. We call these shows Built For You. The last one was chaotic and amazing. We had 5k people live in the comments, cracking jokes and sharing love. If you're an @intercom fan, or just Intercom curious, I think you'll enjoy. The next one in live on LinkedIn in two days! inter.com/bfy-jun
GIF
English
3
23
70
24.5K
Pedro Tabacof retweetledi
Steven Kaas
Steven Kaas@stevenkaas·
You are not the king of your brain. You are the creepy guy standing next to the king going "a most judicious choice, sire".
English
40
4.9K
22.6K
0
Pedro Tabacof
Pedro Tabacof@PedroTabacof·
@reza_kumara @fergal_reid @eoghan @intercom @destraynor Hey Reza, I am part of the Fin Chatbot ML team. This week we shipped an improvement to Fin which makes it more concise (-17% direct answer words) and more readable (+43% paragraph breaks, +133% usage of bullet points). We also made some UI improvements:
Pedro Tabacof tweet media
English
1
0
3
258
Reza
Reza@reza_kumara·
@fergal_reid @eoghan @intercom @destraynor Thanks for the info! Attaching an example here where response is too long ~ 2 paragraphs. While longer answers can be useful, it might be better if users could choose this in Fin AI settings. For me, briefer conversations are more likely to lead to a confirmed resolution.
Reza tweet media
English
1
0
2
354
Eoghan McCabe
Eoghan McCabe@eoghan·
I announced internally last night that we’re going to invest an extra ~$100M to bring a new AI technology we’re dreaming about to market that we’re calling Fin X. Running a scaled company on your own cashflows brings wonderful discipline, and when I returned to @intercom in late 2022 I prioritized getting profitable. But now that we’ve built up our cash reserves, and given the intense competition and massive opportunity in the space of AI for customer service, it’s obvious that now is the time to go way harder to maintain our lead and put that cash to work. Here’s the email I sent with all the context and numbers and internal Intercom jargon. :)
Eoghan McCabe tweet mediaEoghan McCabe tweet mediaEoghan McCabe tweet mediaEoghan McCabe tweet media
English
46
31
582
257.5K
Pedro Tabacof
Pedro Tabacof@PedroTabacof·
@BlakeFolgado @fergal_reid @eoghan @intercom @destraynor Hey Blake, I am part of the Fin Chatbot ML team. Totally agree with you. This week we shipped an improvement to Fin which makes it more concise (-17% direct answer words) and more readable (+43% paragraph breaks, +133% usage of bullet points). We also made some UI improvements:
Pedro Tabacof tweet media
English
0
0
2
208
Blake
Blake@BlakeFolgado·
Becoming more common to see @intercom fin used across products. Responses are super lengthy and mostly unstructured word walls. No human would reply like this so it feels very un-natural, fyi @eoghan
English
2
0
0
930