eshmu

Co-founder and CEO @anomalo_hq. Product and Growth leader @Instacart @Wealthfront @LinkedIn and more. Opinions are my own.

[eshmu @eshmu]

This is literally the #1 thing I learned after years of studying economics (including with @JustinWolfers ) -- Do not mess with the market price!

[Justin Wolfers @JustinWolfers]

If there's a water shortage, don't subsidize showers. If there's a blackout, don't make AC cheaper. If there's an oil shortage, don't cut gas taxes. High prices are the signal telling us to conserve — blunting that signal makes the problem worse.

[Aakash Gupta @aakashgupta]

Tableside Caesar exists for one reason: the dressing has a 10-minute chemistry window. After that, you're not really eating a Caesar anymore.

Crushed garlic releases allicin, the compound that creates the sharp pungent bite Caesar is built around. Allicin peaks 1-10 minutes after the cell walls rupture, then degrades into duller sulfur compounds. Pre-made Caesar sitting in a walk-in for 12 hours has zero allicin left. That's why every bottled version tastes flat compared to one assembled in front of you.

The egg yolk does the heavy structural work. Lecithin molecules wrap individual oil droplets and suspend them in lemon juice acid. A single yolk can emulsify well over a cup of oil. The dressing has body because the emulsion is intact. Once it breaks, you get pooling oil and watery acid. Most pre-made versions use xanthan gum or modified starch to fake the texture, which is why bottled Caesar coats your tongue differently.

Then the umami stack: anchovies, parmesan, and Worcestershire all deliver free glutamates. Three sources of the same savory amino acid hitting your tongue at once. Caesar is the only common salad that activates umami receptors at a level closer to a steak than a plate of greens. The reason it pairs so well with red meat is the dressing is already speaking the same chemical language.

Cold lettuce, warm dressing. The temperature contrast fires the trigeminal nerve before flavor processing even starts, which is why a tableside Caesar registers as more "alive" than one that came from the kitchen 8 minutes ago.

Caesar Cardini invented this in Tijuana in 1924 during a Fourth of July rush when his kitchen ran low on ingredients. He assembled it tableside because the dressing degrades in real time and there was no way to plate it in advance.

Every steakhouse charging $21 for one is selling you a 10-minute window of chemistry that nobody has figured out how to bottle.

[eshmu @eshmu]

Check out this great demo of @anomalo_hq 's new AI Agents from yesterday: https://t.co/uvWR6zasGN

[eshmu @eshmu]

In 20 minutes, come see a live demo of the latest from @anomalo_hq -- our new Agentic Platform with 9 autonomous AI agents to handle common data team tasks. https://t.co/J6NZRQpyw9

[eshmu @eshmu]

@joshelman The best schools now have technology / CS programs. Finance definitely still missing and Economics programs are too theoretical to fill most
important gap.

[eshmu @eshmu]

@villi The rock stars go to John Varvatos (or at least they did back when I last got a suit).

[eshmu @eshmu]

Fun fact: Why is 617 the area code for the Boston area? Because 6/17 is the date of the Battle of Bunker Hill: https://t.co/s3996WqsWU

[Sean Kerner @TechJournalist]

Data quality matters now, moreso than ever. Great to be reminded of that simple fact by @eshmu , whose company @anomalo_hq is out today with new unstructured data quality monitoring tech (and some new $$ too). https://t.co/e0SNCb3WUc via @VentureBeat

[Patrick Collison @patrickc]

Interviewed Alain Bertaud at Stripe this morning, and he said something that I quite enjoyed (paraphrasing): "Part of the purpose of innovation is to avoid the Solzhenitsyn experience in The Life of Denisovich... where every day is predictable. That is the definition of inhumanity. As humans, we need change." I hadn't really thought about the idea of non-predictability as a positive good.

[eshmu @eshmu]

Must say... this is a cool AI demo ... My kids definitely want the Minecraft Copilot.

[Mustafa Suleyman @mustafasuleyman]

We are taking Copilot to the next level. 🚀

Copilot will see, hear, speak and help in real time.

Watch this demo to see what I mean. Soon your AI companion will start to live life alongside you, whether playing Minecraft or helping you navigate life’s most difficult challenges.

Say hello to the future.

[eshmu @eshmu]

@percyliang @stanfordnlp Llama 3 coming soon?

[eshmu @eshmu]

Yes -->

[Aaron Levie @levie]

One of the most common concerns about AI is the risk that it takes a meaningful portion of jobs that humans currently do, leading to major economic dislocation. Often these headlines come out of economic studies that look at various job functions and estimate the impact that AI could have on these roles, and then extrapolates the resulting labor impact. What these reports generally get wrong is the analysis is done in a vacuum, explicitly ignoring the decisions that companies actually make when presented with productivity gains introduced by a new technology -- especially given the competitive nature of most industries.

The thinking generally goes that if a company could, say, be 50% more productive in a particular function, it would mean a commensurate reduction of jobs in that area. For instance, if a certain function (like engineering or sales) required 10 units of labor before, then with a 50% gain in productivity, in the future that same function would now only need ~7 units of labor. The challenge with this type of thinking is that it assumes that companies have maximized the amount of labor they wish they had for a particular function, when in reality many functions are only staffed at the level the company can afford. Further, it assumes that a company is not in a competitive field, and that the company would be complacent and happy about generating the same output as before, just with less costs. Finally, it ignores the fact that productivity gains in a market will lead to increased response from competition, which companies equally have to respond to with more productivity not necessarily more profit. Time and time again this is the type of flawed thinking that we tend to get out of broad economic studies on the labor needs in the economy.

To break this down and make it practical, I thought I'd illustrate the point with the example of an engineering function -- one that already is seeing the benefits of AI starting to roll out. The numbers will all be kept simple, but you can change almost any variable and the point will remain the same. The key to thinking through job impacts is to think through what happens a step or two *after* the productivity gain of AI is experienced.

So, imagine you're a software company that can afford to employee 10 engineers based on your current revenue. By default, those 10 engineers produce a certain amount of output of product that you then sell to customers. If you're like almost any company on the planet, the list of things your customers want from your product far exceeds your ability to deliver those features any time soon with those 10 engineers. But the challenge, again, is that you can only afford those 10 engineers at today's revenue level. So, you decide to implement AI, and the absolute best case scenario happens: each engineer becomes magically 50% more productive. Overnight, you now have the equivalent of 15 engineers working in your company, for the previous cost of 10.

Finally, you can now build the next set of things on your product roadmap that your customers have been asking for. We can't assume it will be 50% more because there are new points of friction and coordination tax that emerge as you have 15 equivalent engineers, but let's say your output goes up meaningfully. Assuming you're acting in your best interests as a company, the features you build make your product that much more compelling, which means at some point (sooner or later) they should result in an incremental gain in revenue. Let's be somewhat conservative on what impact these new features will have on your product, but let's say they generate an incremental 10% of revenue over time or keep customers retained at a 10% greater rate (roughly the same financial benefit).

Now let's assess the downstream impact. Firstly, any growth of revenue will often lead to some functions in the business growing as well to support these new customers, which will directly create new jobs. But further, the company now has to decide whether it remains satisfied with its 10 engineers that have the output of 15, or with their incremental revenue should they hire even more engineers to build the *next* set of features that will make them even more compelling to customers. Unless this company is in some rare monopoly position, they likely will want to build the next set of features even faster than the last set to grow even more quickly. This then means AI has caused the company --counterintuitively-- to hire more engineers than before, because the productivity of each engineer is much higher, allowing them to generate more return per engineer, and thus more revenue.

What's interesting is this analogy works similarly for most functions in a business. In sales, if you could make sales reps 10% more productive (i.e. they sell 10% more of your products/services for the same cost), almost every company in the world would prefer to hire even more sales reps, instead of merely banking the incremental profit. That incremental sales productivity again would lead to downstream implications, like the need to deliver more features to customers, and thus more R&D hiring! Even back-office functions that don't as directly tie to revenue growth, often are a bottleneck to growth . If you can reduce the bottleneck -- say lawyers reviewing contracts, or people processing invoices-- cycle time in businesses accelerates, which almost always lets you serve more customers faster or grow more quickly, again letting a company reinvest those dollars.

In the end, when you step out of the vacuum of just the specific productivity gain of a particular job function, and look at how the whole system will adapt and improve due to that productivity gain, a very different picture of AI's impact on jobs will emerge. Yes there will absolutely be changes to what jobs become more or less in demand in the future, but the competitive nature of companies inevitably ensures that across the whole system companies will be focused on leveraging AI to become more productive.

[eshmu @eshmu]

Re: why large token windows for LLMs are great ...

[Ethan Mollick @emollick]

Gemini Pro 1.5 with 1M token window vs. Claude 2.1 with 100k vs. OpenAI GPT-4 with RAG

I uploaded the Great Gatsby with 2 alterations (mentioning an "iphone-in-a-box" and a laser lawnmower)

Gemini nails it (& finds one more thing). Claude does but hallucinates. RAG doesn't work

[eshmu @eshmu]

This Goody-2 chat bot is amazing ...

[Paul Novosad @paulnovosad]

It's Goody-2, the world's most responsible AI.

https://t.co/ZITRgb58PB

[eshmu @eshmu]

Big believer in this point of view on the impact of AI.

[Anish Acharya @illscience]

Introducing the Abundance Agenda.

We believe AI will turn luxuries into commodities - making consumers happier, healthier, and more productive than ever before.

And a new generation of AI-native companies will lead the way.

More from @a16z consumer 👇

[Steve Stewart-Williams @SteveStuWill]

Self-control is an all-purpose good like IQ: It predicts health, wealth, and all things good. https://t.co/zH6q96TNH8

[SignalFire @SignalFire]

It’s crazy that more data was created in the week it took to write this blog post than in all of 2011. As AI explodes, it’s obvious that we need better tools to keep clean and organized. https://t.co/jlIx9mfGvN

Despite enterprises spending $66 BILLION last year on data infrastructure, errors still run rampant and quality control is the #1 bottleneck to further AI adoption. But AI will also help solve this problem. The first generation of data hygiene was rules-based, and the second focused on metrics-driven data observability. But a company called @anomalo_hq is pioneering third-gen AI-based data quality tools that can sniff out red flag data points that indicate larger accuracy issues.

That’s why we're excited to share that @SignalFire is leading Anomalo’s $33M Series B, alongside @Databricks, @NorwestVP, @FoundationCap, and @TwoSigmaVC to fix data anomalies for the world’s largest organizations. We’re bringing its data luminary founders @eshmu (former Chief Growth Officer of @Instacart), and @jeremystan (former VP Data Science of Instacart) together with SignalFire Executive-in-Residence @douglasmerritt (former CEO of @Splunk), and SignalFire Partner Chris Scoggins (who scaled DataLogix to its sale to @Oracle). Here’s more on how they’re building the future of #AI #data quality at scale.

[eshmu @eshmu]

Fun times with the @anomalo_hq NYC crew this morning! A huge thanks to all the Anomallamas who made this possible.

[Anomalo @anomalo_hq]

🚀 Anomalo secures $33M Series B funding to bring AI-based data quality monitoring to every enterprise 🎉💻

Huge thanks to our amazing team, investors, customers, and partners for making this possible.

📈🔍 #SeriesBFunding #EnterpriseAI #AnomalyDetection #DataQuality #DataOps