Shopify's. CEO says, no new hires till you prove AI can't do the job. Affordable laptops are pulled from sale due to tariffs. The Social Security website in the US crashes and it's linked to Doge software updates and did Meta fudge its benchmark numbers for a new AI version. Welcome to Hashtag Trending. I'm your host, Jim Love. Let's get into it.
Shopify's, CEO Tobi Lutke has issued a directive requiring teams demonstrate that artificial intelligence can't perform a task before they are allowed to seek additional headcount. In an internal memo titled, AI Usage is now Baseline expectation. Lutke emphasized that using AI well is a skill that needs to be carefully learned by. Using it a lot and that AI usage is now a fundamental expectation of everyone at Shopify.
Lutke highlighted the transformative impact of AI on workplace productivity, describing it as the most rapid shift to how work is done that I've ever seen in my career. He said that before asking for more headcount and resources, teams have to demonstrate why they can't get what they want done. Using AI to ensure accountability, AI usage will be incorporated into performance and peer reviews. Shopify has already reduced its head count significantly.
Total head count felt to 8,100 at the end of December, from 8,300 a year before according to its latest annual filing. The Canadian company eliminated 14% of its workforce in 2022 and 20% the following year. The company has also acquired AI focused startups such as Vantage Discovery to enhance its technological capabilities. They're offering an AI driven set of products and services for their customers.
Shopify introduced this suite of AI powered tools for its merchants, including an AI chat bot named Sidekick, and a collection of automation tools called Shopify Magic. But they're clearly trying to take their own medicine. Now, will this spread to other tech companies? Has it already, and if so, will it add to what's already been a large number of reductions in staff at tech companies?
In 2024 alone, there were about 150,000 roles eliminated over 549 tech companies according to layoffs.fyi given a potential economic slowdown. We might even be looking for more companies looking to use AI to drive cost reductions, particularly in the area of headcount reduction.
I. Framework. A San Francisco based laptop manufacturer known for its customizable and repairable devices has temporarily halted US sales of certain base models in its laptop 13 series due to new tariffs imposed by the US government.
The affected models include configurations featuring Intel's Ultra Five and AMD's Ryzen five Processors framework is an interesting company, not just because of their commitment to repairable and upgradeable laptops, but they also have good price points, especially given the fact that these laptops can continually be updated for longer lifespans Based on the long-term value and the environmentally friendly stance of the company, it grew by over 30% last year.
But recent tariff increases may hinder that growth. Its lowest priced laptops with the Intel Core Ultra five and the A MD rise in versions where selling for $999 US and $899 US respectively before the tariffs. Now with the lowest tariffs, 10% on Taiwanese shipments, these machines would have to be sold at a loss. So the company has pulled them for now without these models, the next cheapest versions of those laptops will start at $1,399 US and $1,499 US respectively.
Framework is certainly not the only company suffering under these tariffs. The company's commitment to transparency puts them out ahead of others who are certainly facing similar challenges, but who might not yet have publicly disclosed their actions. Expect more price adjustments and less availability, especially of affordable lower end models. And Meta is facing criticism for using an internal version of its Maverick AI model in benchmark tests without clearly disclosing it.
The version optimized for chat helped Maverick score 1,417 on LM Arena, ranking it above open AI's GPT-4 oh, and just behind Gemini's 2.5 Pro. I know the names are confusing, but Maverick is actually part of the LAMA four release. It is a relatively compact model and uses a mixture of experts and reinforcement learning to help it excel at complex tasks like coding and other things.
However, LM Arena says that in the benchmarks meta used an experimental chat version that differs from the public release, which raises concerns about the fairness and transparency of the results. The benchmarking site faulted meta for not clearly distinguishing between the internal and the public models. I. But some researchers have also alleged that meta may have trained its model on parts of the test data. Meta's, VP of generative AI denies those claims.
But the lack of transparency has sparked renewed debate over benchmark integrity in the AI industry. LM Marina is an open source platform for crowdsourced AI benchmarking. It was developed by researchers at the UC, Berkeley Sky Lab, and it's designed to evaluate and rank large language models, LLMs and AI chatbots based on human preferences through anonymous randomized comparisons.
Now the anonymous nature of the ratings and the fact that they're not pure numbers, but reflect the actual experience of users with the software have made LM Arena a way to cut through the dozens of different benchmarks and give what was hoped to be a realistic view of the user experience. So while Meta didn't technically break any LM Arena rules the incident affects the trust, not of the benchmarks, but of the AI companies themselves.
And the Social Security Administration, the SSA in the US has experienced multiple website outages in recent weeks, disrupting services for beneficiaries. These issues have been traced to a recent software update mandated by the Department of Government Efficiency, or Doge, led by Elon Musk. The update intended to enhance fraud detection was implemented without adequate testing for the higher user volumes that would result leading to system crashes.
DOGE initiative expanded existing contracts with credit reporting agencies to perform earlier and more extensive identity verifications during the claims process. However, the SSA's technology staff did not conduct sufficient load testing resulting in connectivity issues and authentication portal failures. Now compounding the problem, the SSA site already had issues and Doge directed budget cuts had led to a significant staff reduction in the SSA, particularly within the IT department.
The result has been a perfect storm. With reductions, forcing more users to rely on the malfunctioning website with fewer employees available for phone support. And in turn there are far fewer IT resources to help solve system issues on some of these ancient systems. , the inevitable result is service disruption on what for many is a major portion of their retirement income.
This also coincides with more people having to come into Social Security offices to answer issues from these new questions on their claims. Although there are not enough staff to serve the new demand and scheduling systems are malfunctioning, but don't worry, rumors are that Doge is going to redo the Social security system and.
They say it will only take a matter of months now that they've gotten rid of a lot of the internal IT talent and that pesky knowledge of the old system, what could possibly go wrong? It might just be me, but there are people who need to be let go. They're just not the social security IT staff. And that's our show for today. You can reach out to me at [email protected] or on LinkedIn or if you're watching on YouTube, just leave me a comment under the video.
I'm your host, Jim Love. Have a wonderful Wednesday.