Desteia (formerly Auba)

The future of supply chain? The future of work? AI isn��t the big story- this is. I was at Manifest (future of supply chain conference) in Las Vegas last week. I thought everyone would be talking about AI agents revolutionizing logistics (boring). That��s not the real shift happening. After dozens of 1:1 conversations with supply chain execs at Manifest (future of supply chain conference) last week, I realized the real future is something else entirely��something most AI startups aren��t thinking about. And if you don��t get ahead of it, your projects (and maybe your career) could be dead in the water. Here��s what I learned.?? 1. Storytelling is the premier skill to develop. - Executives are over-saturated with ��data��. Tableau reports. Industry reports. Data all around. AI all around.? - Thus, the important skill is: can you tell a story with the data? Can you convince an executive to sponsor your project based on that data? Can you help companies wrap their arms around that abundance of data? Can you explain something complex, easily? - If you can��t, your project is dead. In an age where AI can outperform any analyst, being able to tell a convincing story will become the #1 skill. 2. Data Science everywhere.? - This one blew my mind. 5 years ago, close to 0% of job descriptions had a component of data science in them.? - Today, close to 20%.? - Mix this with my first point on story telling and a clear picture emerges: the future belongs to those who can develop insights, tell a convincing story through data, deal with change management, and use AI effectively. 3. ��Being ready�� for AI is BS ... there's not enough characters to post all my thoughts here. Please continue reading in the article link shared below. If you like this post, please help me share it and comment, would love to hear your thoughts!

Make sure to listen to our latest podcast #supplychainoptimizers

As the supply chain and logistics world gathers this week for Manifest: The Future of Supply Chain & Logistics, the latest edition of "Between Two Lightbulbs" focuses on our new portfolio company, Desteia (formerly Auba), and its co-founder, Diego Solorzano. Desteia is tackling one of the industry��s biggest challenges��supply chain visibility��with an AI-driven platform that delivers real-time shipment tracking, predictive analytics, and proactive risk management. As volatility continues to disrupt global logistics, particularly in North America, Desteia empowers businesses to stay ahead with smarter, faster decision-making. Check out the video, and if you��re at the conference this week, let��s connect!

Sharing these news today! if you're looking to make your US-MX supply chain faster, cheaper, and easier to manage ping us!

There's a silent killer shutting down most supply chain optimization projects. Controlling it, is one of the keys to Amazon's success. That��s what I learned in our most recent episode of Supply Chain Optimizers chatting with Carlos Armando Zetina, Ph.D. Carlos is an incredible academic turned operator. Most recently, he��s been a part of Amazon Supply Chain Optimization Technologies (SCOT) working to optimize Amazon's supply chain in real time. Here��s what I learned from him this week: 1. ???????????????????? ???? ?????? ?????????? ???????????????? ???????? ???????? ???????????? ?????????? ???????????????????????? ???????????????? (????????????, ???????? ?????????????????? ????????????????) Academia is designed in such a way that complexity is a highly valuable possession. Academics are able to make complicated models with extremely sophisticated systems. They need to in order to publish. In the real world, these models tend to be difficult to replicate. A good rule of thumb: If it takes more than five minutes to explain, it is too complex to be implemented. Even If people get around to understanding the model, most wont trust it due to its complexity. 2. ???????????? ?????? ??????????; ?????? ???????? ?????? ???????? ????????????. Models often fail to replicate the world exactly as it is. Still, they are a great approximation that can save us a lot of time. The problem (as in point #1) is when we believe a model can do everything. A simple model, to address a specific pain point will be more successful than most "catch-all" models sold today. 3. ??????'?? ?????????? ???????? ????????????. A lot of CEOs ask their teams to be "more like Amazon". This is wrong. It ignores the fact that Amazon has a very specific business model (the everything store for everyone everywhere), and decisions (and models, and projects) are made for that business model. If that's not your business model, there's no point in "trying to be more like Amazon". Focus on your specific business model and works for it. 4. ???????????????????? ???? ?? ?? ?????? ???? ?? ???????????????? ???? ???????????????? ???????? ?????????????? ???????? ?????? ?????????? No such thin as "hands off": most companies dream of having operations without any human interaction. But machines are far from perfect and lack context from the world. We humans are constantly reading the news and talking with our peers��giving us an edge when it comes to adjusting to unforeseen circumstances. Automation is great in some settings. But humans still have more information for key strategic decisions. you need humans to make key strategic decisions in some settings at least. Thanks Carlos!! ---------------------------------------------------------------------- Let us know what you think about this episode and who you'd like to see on next as a guest! And, if you're looking to save your company money through better supply chain decisions, hit me up or check Desteia (formerly Auba)

Everything you need to know about Deepseek, the US-China AI race, and the golden age of AI, in less than 60 seconds. If you follow AI news, you surely read about Deepseek this weekend: A Chinese startup that released two very impressive models (v3- their gpt4 equivalent, and r1- their o1 equivalent). Here's the kicker: they made it fully open source and (apparently) trained them with 1/10 the cost of other LLMs from the likes of OpenAI, Meta and Google. This has MAJOR (MAJOR) implications for the AI world and probably is the biggest AI moment since the launch of ChatGPT in 2022. Here��s what you need to know: I won��t go into the conspiracy theories, you can find plenty of those online. The one thing that does strike me as a big coincidence is that these models were released on President Trump��s first week in office, as this release could be interpreted as China��s gift to the world. ???????? ???? ???????? ???????? ???? ?????? ??????????? 1. Showing that it��s possible to train world class models for under $100M 2. Making it fully open source (while performing on par or above its closed source equivalents form OpenAI and Anthropic) 3. Opening the AI race (popping the bubble?) wide open for every country in the world, including Mexico. Let��s break these further down. 1. It��s possible to train world class models for under $100M (and A LOT of data) Up until this release, the world was following OpenAI��s lead. OpenAI had a simple message: ��all you need is compute���Cie, more machines to train AI models (called GPUs). According to OpenAI��s theory, the more GPUs, the better the model. Thus, the assumption goes, to build AGI (or ASI, the holy grail of AI systems) you need massive, ever expanding, levels of compute. That��s why Sam Altman and president Trump announced project Stargate last week (a $500bn investment in GPU clusters). With Deepseek reportedly costing a little over $50M to train, the thesis above is now heavily questionable. (Thus why Nvidia stock is dropping today). And, if such large clusters are perhaps not required, then the energy to feed them might not be required either (so, far less environmental impact). Word of caution: access to data could still be a bottleneck. If compute was king, then data is queen. The king is dead, long live the queen! 1. 5 From the age of training to the age of inference? I read through Deepseek��s papers and a bunch of posts on this so you don��t have to... ... there's not enough characters to post all my thoughts here. Please continue reading in the article link shared below. If you like this post, please help me share it and comment, would love to hear your thoughts!

Excited to launch our podcast ��Supply Chain Optimizers��! This week: Jack Levis, the man who saved UPS more than $400 million yearly by banning* drivers�� left turns- amongst other optimizations. Here's what I learned. ???? ?????????? ???????? ???????? ?????? ???????????????????? ???? ???????? ???????????? ???????? - Change management is the #1 reason why projects fail. - You need buy-in from management as much as you need controls, tests, pilots, and measurements. - You also need buy-in from users. Jack��s projects would��ve failed if UPS drivers did not adopt them as their own. There are ways to do this (hint: it's what you measure, stupid) ???? ?????? ??????��?? ?????? ��???? ???????? ???? ???? ?????????????????? ???? ???? ??????????????��, ?????? ???????? ????????. - Every optimization and savings project is based on a strong data backbone. Make gathering quality data a core competency of your business. - The process is: good data �� advanced analytics �� optimization. ??????��?? ?????? ?????????????????????? ???? ?????? ?????????? ???? ????????????????. ?????? ?????????????? ???????????? ???????? ?????? ?????? ??????????????????????????. ?????? ???????????? ???????????? ?????? ????. - There��s more ways for a single UPS driver to make 120 deliveries in a day than there are atoms in the universe. Read that again. - A lot of times, for optimization, it��s the small things done consistently that add up to the big savings (avoid unnecessary left-hand turns, have your pen in your left pocket if you��re right-handed, where to hold your keys, don��t leave your vehicle idle). ????��?? ?????? ???? ?????? ??????????????: ??????????????, ??????????, ?????????????? ?????????????????? ???????????????? ???? ??????????????. ??????????????, ??????????????, ??????????????. - For UPS, 1 mile saved per driver per day = $50M per year. - 1 minute saved per delivery per day = $14.5M. - Reducing vehicle idle time 1 minute per day saves $500k in gas per year. ?????????????? ?????????????? ???? ???????????? ???????????????????? ?????? ???????????????????? - Data is what will run your company. - Virtual and physical worlds need to be aligned. - Listen to operators: are they adopting and making your projects their own? Are they competing for the most savings or ��best driver��? - Easiest way to change behavior: change what you measure. Measure what you can control (behavior), not the savings. The savings will follow. * per Jack, the right way to say it is ��avoid unnecessary left hand turns��. Thanks for all of your insights Jack Levis! ---------------------------------------------------------- I��m excited to launch our podcast ��Supply Chain Optimizers�� by Desteia (formerly Auba) where we interview supply chain, logistics, and optimization leaders about real-life savings and innovations. Hopefully, we can inspire you to take action in your company! What do you think? Leave a comment and DM me if you��re interested in exploring potential opportunities for your supply chain!

??? Big Update: AUBA is now Desteia??? After much thought and planning, AUBA is evolving to Desteia. This change is not merely a new name; it represents our deepened commitment to advancing supply chain efficiency, visibility, and innovation. The name Desteia is inspired by the Spanish word ��destello,�� meaning a ray of light, symbolizing clarity, speed, and illumination��values which are close to our own vision for logistics. With Desteia, you can expect the same dedication, solutions, and excellence you've come to rely on, now with a renewed focus on enhancing our platform and expanding our capabilities. Our mission remains unwavering: to give logistics leaders the tools and insights they need to make informed, strategic decisions. So, without further ado, join us in welcoming Desteia to the world!?(edited)?

?? It's been quite a week for supply chains! Here's everything you need to know with our newsletter ??

I spoke with 33 logistics leaders in the last two weeks, and something most of them ignored really shocked me: They didn��t understand that (probably) 97% of their shipping routes rely on transhipments. As always, we crunched the numbers at Auba, here��s what we learned. Only 2.9% of global port routes are direct connections. That means the vast majority��97%��of global shipping routes depend on transhipments. Yet, many supply chain professionals underestimate their impact. These are potential choke points that can shipping timelines. The numbers: * 939 container ports exist globally, but only 12,748 direct routes connect them. If every port had direct connections, there would be 440,391 routes��but that��s far from reality. * A shipment may require up to 6 transhipments and 14 port stops to reach its destination, with every stop increasing the risk of delays. * A disruption at one port can trigger a logistics waterfall effect, causing delays of days or even weeks. In a previous post, we estimated the effect to be over 32 days for Mexican ports. The Port Liner Shipping Connectivity Index (PLSCI) ranks ports by their connections and throughput (check out the cool map below). Here��s what the data shows: * Shanghai dominates globally with a score of 2,258.52, far ahead of any other port. * Global average connectivity has risen from 89.26 (Q1 2006) to 101.4 (Q3 2023)��a step forward, but regional gaps remain: U.S.: 126.56 points (highest in North America). Mexico: 94.11 points (second, outperforming Latin America��s average of 92.27). Canada: 74.81 points (20 points behind Mexico and 50 behind the U.S.). So what? Transhipments don��t just delay individual shipments��they can create bottlenecks across entire regions. A delay at a single port can delay hundreds of supply chains. ---------------------- The Opportunity to Optimize This isn��t just a problem��it��s an opportunity. With the right data, you can identify high-connectivity ports and model cascading delays across 6+ transhipments or 14 port stops to plan proactively. Imagine how much stronger your supply chain could be if you built a strategy rooted in the numbers��not just guessing where delays might happen, but knowing exactly where they will.