A week's worth of tech news that takes you 5 minutes to read

Welcome to this week's Sunday Download! What we do is simple: give you a week's worth of tech news in 5 minutes or less so you can stay informed and up to date on the latest technology news and trends.

We cover everything (or at least try to) from Big Tech, mobility, AI and Machine learning, fintech, futuristic tech, and everything in between.

In return, we ask that you share it with a friend or colleague instead of keeping the Smartology goodness all to yourself. While greed may be good on Wall Street, sharing is caring here at Smartology. 🤝

Lots of stuff to go through this week, so let's not waste any (more) time. Highlights of this week's issue include:

• Cellphone spying without a warrant

• AI, API, ChatGPT, and other acronyms

• Salesforce fights back

Total read time: 4 minutes and 32 seconds. Let's goooooo! 🚀

No Warrant? No Problem! (Actually, a big problem)

According to a government watchdog, the Secret Service and ICE's Homeland Security Investigations (HSI) section routinely failed to secure the proper legal papers when conducting invasive mobile phone surveillance. Meaning: They didn't get the warrants. 🤦 Apparently, this has been going on for several years.  

The inspector general of Homeland Security announced the findings last week, alleging that the agencies frequently utilized cell-site simulators (also called Stingrays) without getting the necessary search warrants.

You can watch a quick video about how Stingrays work here, but, to summarize, they mimic cell phone towers and can gather data on everyone in the area they cover. Phone call history, texts sent, and location data, to name a few. The problem is that the Stingray targets everyone in its geographical zone (civilian or criminal). As a result, federal agencies must first get a search order from a court before using a cell-site simulator, which the Secret Service and HSI did not do. 🫤

ChatGPT goes API

On Monday, OpenAI stated on the company's blog that it provided APIs for ChatGPT and Whisper (its lesser-known brother with a phenomenal speech-to-text model). This enables developers to create new applications or add features to current ones without starting from scratch.

Developers can now incorporate the ChatGPT API into their products to construct chatbots, virtual assistants, and other applications without having to train and host their own language models.

The potential extends far beyond chatbots, as the technology will almost certainly lead the way for an era of virtual assistants, natural-language coding, and automated marketing.

ChatGPT's ability to have natural-language conversations with humans swept the globe in late 2022, bringing in more than 100 million active users since its launch. 

And not terrifying at all is this article which shows that ChatGPT can now control physical robots through its API offering. 😱 I keep warning people about the robot takeover coming faster than we think, but no one seems to listen. 

Tiktok has more problems

TikTok can't seem to get away from problems. Earlier last year, a number of educational institutions prohibited the use of TikTok on mobile devices. A week ago, the European Commission banned using TikTok on work-related devices. This week, Canada's banned the use of the app on government devices, following in the footsteps of the European Commission. The driver for all of this is that the China-based TikTok platform can be used to spy on its users.

The US is presumably next to implement some version of a TikTok ban, but how easy would that actually be? The answer: near impossible. This article does a great job of discussing why a national ban on Tiktok would be easier said than done. 

Dish Network Outage

Dish Network (who also owns Boost Mobile) got hit with a ransomware attack this week, which took down their TV service for almost five days. The outage impacted nearly 8 million subscribers. 

The system issues started on 2/23, and they're still not up and running. Here's what you see when you go to dish.com: 

Dish has no clue who gained access to their systems, what they got access to, and how bad the damage is. The official corporate line is "We're making progress on the customer service front every day, including ramping up our call capacity, but it will take a little time before things are fully restored." Read more here. 

Meta joins the AI race

Meta will establish a new product group devoted to machine learning and AI. The move comes as large technology corporations and well-capitalized startups compete to showcase improvements in machine learning techniques and incorporate artificial intelligence models into their products. In addition, the new department will create "creative and expressive" tools to be used within Meta's products.

This week, Meta also announced the release of LLaMA, a new large language model that is smaller and less expensive than larger models. The announcement indicated that Meta, which conducts considerable artificial intelligence research (and has been doing so for some time), would not let competitors pass it by in the AI race for our hearts and wallets. 🫶

FTC: Keep your AI claims in check

The Federal Trade Commission (FTC) warned companies looking to cash in on the recent wave of AI madness, telling them to "keep your AI claims in check." There will be a short leash for those writing checks that their tech can't cash. 

Questions companies need to ask themselves (according to TechCrunch):

• Are you exaggerating what your AI product can do? If you’re making science fiction claims that the product can’t back up, you may want to tone those down.

• Are you promising that your AI product does something better than a non-AI product? Sure, you can make those weird claims like “4 out of 5 dentists prefer” your AI-powered toothbrush, but you’d better have all 4 of them on the record. Claiming superiority because of your AI needs proof, “and if such proof is impossible to get, then don’t make the claim.”

• Are you aware of the risks? “Reasonably foreseeable risks and impact” sounds a bit hazy, but your lawyers can help you understand why you shouldn’t push the envelope here. If your product doesn’t work or its results are biased because your dataset was poorly constructed… you’ll have a bad time. “And you can’t say you’re not responsible because that technology is a ‘black box’ you can’t understand or didn’t know how to test,” the FTC adds. If you don’t understand and can’t test it, why are you offering it, let alone advertising it?

• Does the product actually use AI at all? Claims that something is “AI-powered” because one engineer used an ML-based tool to optimize a curve or something doesn’t mean your product uses AI, yet plenty seem to think that a drop of AI means the whole bucket is full of it. The FTC thinks otherwise.

Salesforce gets a win

It would be an understatement to say that Salesforce has been having a rough ride of late, facing massive activist investor pressure and an exodus of its executives. Marc Benioff has not been having a good time.

This week, Salesforce posted fiscal fourth-quarter earnings, including much-needed reports that their revenue and guidance exceeded market expectations. With that, Salesforce announced it would increase its share repurchase program to $20 billion. This is in response to external criticism of its cost structure and expectations for growth in the current fiscal year. The firm intends to buy back shares and has the cash flow to do so, and its earnings report has proven its opponents incorrect and given it more time to prove that it knows what it is doing.

The hottest job in tech: Prompt Engineer

Knowing how to talk to chatbots is the undisputed must-have skill of the future. Klarity, an AI contract review firm, is looking for an engineer to "prompt, fine-tune," and "chat with" large language models for up to $230,000 a year.

So what's a Prompt Engineer? Prompt Engineers are experts in asking AI chatbots, which run on large language models, questions that can produce desired responses. Unlike traditional computer engineers who code, Prompt Engineers write prose to test AI systems for quirks, which is required to develop and improve the human-machine interaction element in the software. 

One of the biggest criticisms of AI chat programs is that they have a tendency to be comically inaccurate (check out the hilarious and insightful Twitter feed of Riley Goodside, a Prompt Engineer at Scale AI). This is where Prompt Engineers can help by systematically teaching the AI system how to find the correct answer, ensuring accuracy and safety for all involved. 

Hydrogen Fuel is closer to reality

A Universal Hydrogen-branded plane, equipped with the largest hydrogen fuel cell ever to power an aircraft, made its maiden test flight in eastern Washington. It flew for 15 minutes and 15 minutes and showed that hydrogen could be viable as a fuel for short-hop passenger aircraft. The plane used was a Dash-8, a staple at regional airports, and can usually transport up to 50 passengers on short hops.

Rapid Fire 

• When is 6G coming? Spoiler alert: Not for a while

• Snapchat launches "My AI" and it's kinda creepy

• Tesla to open a factory in Mexico

• VW wants to put apps in your car

• Four Experts on Tech Hiring "Where are the jobs?"

• New app summarizes long emails

• DoorDash launches a credit card with Chase

• AR headset helps give people x-ray vision

• Is this the cure for the hiccups?

Are you interested in sponsoring this newsletter? If so, send an email to [email protected] to find out more!