Contrary Research Rundown #109
CT scans for factories: how Lumafield plans to use AI to reinvent manufacturing, plus new memos on MoveWorks and Cointracker.
Research Rundown
Although our lives are increasingly online, we still live in a physical world where the limits of technological innovations are determined by the quality of physical devices, and the manufacturing capacity of the economy determines the price of goods.
Failures in quality control for physical devices can be costly. For example, the 2016 Samsung recall of 2.5 million units of the Galaxy Note 7 cost $5.3 billion because the batteries were heating up and causing fires. Tesla had to recall all Cybertrucks manufactured between November 2023 and April 2024 due to a defect where the accelerator pedal could get trapped inside the vehicle’s trim, putting drivers’ lives at risk. A production breakdown at a Boeing factory caused the door of Alaska Airlines Flight 1282 to blow off mid-flight at 16K feet; had it occurred at cruising altitude, passengers likely would have died.
However, the increasing quality of AI presents an opportunity to improve quality control and the manufacturing process as a whole by helping companies more effectively identify and correct defects before products are pushed through production. Lumafield wants to address this opportunity by building a high-quality manufacturing dataset through CT scans in order to increase the accuracy and reduce the costs of industrial inspection and, eventually, to enable fully autonomous manufacturing.
X-ray computed tomography (CT) scans, which you may have heard referred to as “CAT” scans at a doctor’s office, take multiple images from different angles to construct a 3D model of their target. In medicine, this is useful for helping doctors diagnose tears and breaks. CT scans can be used the same way in manufacturing by helping companies more effectively identify and fix product defects before they are shipped to the public without the need to rely on traditional and potentially destructive testing methods. However, high costs have historically prevented smaller companies from accessing industrial CT scans.
Lumafield provides low-cost industrial CT scanners called Neptune and Triton, as well as a browser-based CT analysis system called Voyager, to allow companies to affordably inspect and improve the quality of their manufacturing. However, local CT imaging solutions are only the first step. In a September 2024 interview with Contrary Research, Lumafield’s CEO Eduardo Torrealba explained that Lumafield’s larger vision was to use data across many scans to enable fully automated manufacturing processes.
At present, a broad, high-quality database of manufacturing information doesn’t exist online, making it difficult to train AI to identify and resolve product defects. However, with time, Lumafield wants to build that database from its scans. As it collects more data points from customers, its Atlas AI agent will grow more effective at identifying product anomalies and expediting quality assurance and manufacturing processes.
To learn more about Lumafield and how the company is addressing this quality control opportunity, check out our new memo on the company, as well as our recent conversation with Lumafield’s CEO, Eduardo Torrealba on Research Radio.
Moveworks was founded with the intention of enabling employees to get their questions answered faster. Its AI Copilot integrates with dozens of apps that employees already use like Slack and Microsoft Teams so they can get their questions answered instantly, without interrupting their workflow. To learn more, read our full memo here and check out some open roles below:
Senior DevOps Engineer - Remote, US
Senior Network Architect II - Mountain View, CA
CoinTracker enables consumers to track their crypto portfolio across more than 800 exchanges and wallets, helping users see their market value, investment performance, and transactions, and prepare taxes as they transact with cryptocurrency and other Web3 assets. To learn more, read our full memo here and check out some open roles below:
Product Designer - Remote
General Application - Remote
Check out some standout roles from this week.
Found | San Francisco, Seattle, Portland, New York, or Remote - Senior Product Data Scientist, Senior Software Engineer Full-Stack (Risk), Data Engineer
Character.AI | Menlo Park, CA - Head of Design, UX Researcher, Security Engineer (Privacy), Software Engineer (Growth), Research Engineer (Multimodal)
Mercury | San Francisco, New York, Portland, or Remote - VP of Design, Senior Software Engineer (Risk Intelligence), Senior Backend Engineer (Personal Banking)
Anyscale | San Francisco, CA - Research Engineer, Engineering Product Manager (Ray Core), Software Engineer Observability (Backend), Consulting Engineer
OpenAI released ChatGPT Search which offers web search capabilities, intended to provide quick and relevant answers with links to sources. Its goal is to combine the ease of natural language interaction with access to real-time information like sports scores, news, and stock quotes.
A new report from Reuters claimed that Intel CEO Pat Gelsinger offended TSMC with some public comments he made in May 2021, causing the company to end a deal it had with Intel that included a 40% discount. Intel announced 15K layoffs earlier this year, its largest ever.
OpenAI, which is one of the largest buyers of chips, is reportedly working with TSMC and Broadcom to build its first in-house AI chip to secure its chip supply and manage costs.
Sierra, an AI startup co-founded by former co-CEO of Salesforce Bret Taylor, has raised $175 million in funding at a $4.5 billion valuation just one year after its launch. At $20 million ARR, this valuation represents a 225x revenue multiple.
Google is reportedly developing a “computer using” AI agent called Project Jarvis that can carry out tasks for users like gathering research, purchasing a product, or booking a flight.
Neuralink founder Elon Musk stated in an interview that he thinks the company should focus on creating an implant that can eliminate back and neck pain to help improve people’s happiness and enhance the quality of sleep.
Meta is reportedly developing an AI-powered search engine to reduce its reliance on Google and Microsoft, as its Meta AI bot currently embedded in Instagram and Facebook relies on Google for information. This new AI search engine would generate real-time summaries of current events directly within the Meta AI chatbot.
The US is finalizing regulations to prohibit investments in specific AI technologies in China, particularly those linked to military and surveillance applications. Beginning January 2, 2025, US investors will be barred from supporting China’s advancements in AI for military, intelligence, and cybersecurity purposes.
Anduril said that it welcomed a White House memorandum on advancing US leadership in AI, highlighting the urgency of staying ahead of strategic competitors like China in AI when it comes to national security.
Hadrian has appointed West Owens as its new CFO, whose previous role was CFO at TerraWatt. Hadrian CEO Chris Power expressed confidence that under Owens' leadership, the company will expand its business reach and enhance the capabilities it offers to customers.
Productboard announced a new AI-powered feature called Productboard Pulse which is intended to gather insights from various sources, condense them into actionable trends, and align them with users’ business goals
Boston Dynamics has released a new video of its Atlas robot. Previous videos of the all-electric robot showcased its distinctive range of movements, but this time, Atlas has demonstrated that it can utilize machine learning and enhanced sensors to carry out sorting tasks in a simulated factory setting.
According to Reddit’s Q3 shareholder letter, its daily users had increased 47% year over year, partially driven by an increase in traffic from Google, indicating that Reddit is increasingly seen as a source of human information as AI-produced content becomes more prevalent.
At Contrary Research, we’re building the best starting place to understand private tech companies. We can't do it alone, nor would we want to. We focus on bringing together a variety of different perspectives.
That's why we're opening applications for our Research Fellowship. In the past, we've worked with software engineers, product managers, investors, and more. If you're interested in researching and writing about tech companies, apply here!