Gearshift time, Tech Insiders. Voice bots stall at Taco Bell, a cyberattack slams JLR into park, and Benioff's AI hits the gas on layoffs—proof the future still grinds its gears.
Fire up the engine; we'll guide you through the smoke. |
|
|
Here's what you need to know today: |
|
|
Taco Bell Hits AI Drive-Thru Wall |
Hold the bot… add extra human sauce.
Just over a year after boasting its AI would "elevate the order experience," Taco Bell's voice-AI rollout across more than 500 drive-thrus is sputtering after trolls ordered 18,000 cups of water and regular customers got looped into soda-selection purgatory.
Chief Digital and Technology Officer Dane Mathews admits the chain is "still figuring things out," echoing McDonald's earlier retreat from automated ordering, which infamously added bacon to ice cream. Viral clips of the system mishearing orders have racked up tens of millions of views, turning the tech hiccup into brand theater. |
The company says the bot has processed more than 2 million orders, but peak-hour accuracy remains shaky. Humans are being coached to jump in behind the headset at busy locations, turning the futuristic system into a sometimes-on convenience.
Parent Yum! Brands is still buddying up with Nvidia for future AI projects, yet Taco Bell is slowing expansion until it can prove the tech actually delivers on its original promises to trim wait times and errors... without inviting another meme-storm.
Competitors like Wendy's and White Castle face similar friction, suggesting voice AI hasn't cracked the greasy-bag code. For now, the chalupa chatbots are on probation.
Why it matters: If a fast-food giant with a notoriously complex and customizable menu can't tame drive-thru AI, retailers everywhere may pump the brakes on automating frontline customer chatter, especially when a single viral fail can drown out any labor savings. |
|
|
Would you use voice AI to order fast food? |
|
|
Results from Yesterday's Pulse Check |
How much faith do you put in made-in-China AI chips? |
|
|
Rising Token Tabs Threaten AI Startups |
Wallets are screaming louder than the models.
The cost curve for cutting-edge AI is flipping. The price per token is drifting down, but tasks now burn through tens of thousands more tokens as models "think" harder by rerunning queries and checking their own work.
While chatbot basics might gulp 500 tokens, multistep agents can slurp a million, turning cheap-to-meter dreams into budget nightmares. |
Indie devs are feeling the burn. Users of coding assistant Cursor report that their monthly credits now disappear in days. Over at Replit, a new "effort-based pricing" model sparked loud backlash on Reddit, though its CEO insists the company "didn't see any significant churn" and still boasts fat enterprise margins.
Some devs threaten to bail; others scale back prompts or swap to lightweight models. Meanwhile, bigger fish like Notion swallow the hit; CEO Ivan Zhao pegs a 10-point margin dent yet stays profitable thanks to volume.
McKinsey forecasts AI infrastructure spending could hit $6.7 trillion by 2030, favoring giants that own chips, power, and data centers. For startups, survival may hinge on tighter rate limits, a seriously wealthy best friend, or swapping to dumber models, which can be over 30 times cheaper than their genius siblings. If your agent writes code that bankrupts the repo, who's really doing the automating? |
|
|
🎯 Share AI Strategy for a Chance to Win $100 Gift Card Take the State of Industrial AI and Cloud Strategies 2025 survey to benchmark your peers and enter a drawing for two $100 gift cards. 🤑 Score Hot Leads with Pipedrive Pulse
Try Pipedrive Pulse, your AI-powered lead qualification ally, to auto-rank prospects and jump the waitlist before rivals do.
📧 Compare Top Email Marketing Platforms Now Dive into the Best Email Marketing Software and find the tool that lifts open rates without lifting your budget.
💻 This section contains sponsored tech insights. Advertise with us! |
|
|
AI at the Gate: Why Identity Security Starts with Smarter Passwords |
Identity threats are evolving—and passwords are the new battleground.
Join this webinar on September 10 at 11:00 AM ET, hosted by Specops Software and TechnologyAdvice, to explore how smarter password strategies can protect your workforce and brand. |
| |
Hack Stalls JLR Manufacturing Lines |
Adding insult to its already existing financial injuries, Jaguar Land Rover, owned by India's Tata Motors, yanked the plug on global IT systems after detecting a weekend cyberattack.
The hack slammed its UK assembly plants in Solihull and Halewood, froze dealer portals during UK new-plate week, and marked what security experts are calling the automaker's second major breach this year. |
With production halted and no timeline for a reboot, analysts, speaking to Sky News, peg downtime at up to $2.1 million (£1.6 million) per hour.
The company insists no customer data was stolen. Still, the incident follows crippling cyberattacks on Marks & Spencer and the Co-op Group, proving the UK's supply chain is on the naughty hackers' speed dial.
Adding a final layer of irony, the hit comes just two years after JLR inked a $1 trillion (£800 million) cybersecurity deal with fellow Tata firm Tata Consultancy Services. "Zero trust" shouldn't describe your confidence in getting a new Defender this month. |
Google Slams Brakes on 'Massive' Gmail Hack Story |
Google yanked the emergency cord on a viral story claiming 2.5 billion Gmail accounts were at risk, calling the reports "entirely false."
A full-blown panic that had been building for days peaked at the start of the week after a bizarre game of telephone conflated a real (but minor) breach of a corporate Salesforce server with a general phishing warning, sending users scrambling to change their passwords.
Google insists the actual breach only exposed basic, publicly available business info. So, not exactly your diary. The company says its systems still block over 99.9% of phishing attempts, though it did use the chaos as a chance to remind everyone to maybe, finally, use passkeys. Your password is fine. The internet's collective blood pressure, on the other hand, is not. |
'I Need Less Heads': AI Agents Nix 4,000 Salesforce Support Jobs |
Marc Benioff just rehired the robots.
On a recent podcast, the Salesforce CEO revealed customer service headcount dropped from 9,000 to 5,000 after "Agentforce" bots began handling half of all support chats, slashing support costs by 17%. Separately, Benioff boasted that sales AI is finally tackling a 100-million-plus-lead backlog from over 26 years.
A year ago, Benioff swore AI would only augment workers; now he says, "I need less heads." |
The shift mirrors Klarna's AI support, Microsoft's sales layoffs, and Meta's staff culls, signaling that white-collar roles once deemed safe are squarely in AI's crosshairs. However, some analysts say AI is just a convenient excuse for correcting pandemic over-hiring.
Benioff still calls the human-bot mix "exciting," but displaced staff, even the hundreds Salesforce says it redeployed, might choose other adjectives. |
|
|
|
Writer at TechnologyAdvice |
Justin Meyers is an investigative writer and editor who draws on over a decade of meticulous hands-on research to deliver the full, trustworthy story behind consumer and enterprise tech, including cybersecurity. |
|
|
| Writer at TechnologyAdvice |
Justin Meyers is an investigative writer and editor who draws on over a decade of meticulous hands-on research to deliver the full, trustworthy story behind consumer and enterprise tech, including cybersecurity. |
|
|
Curious about where AI is really headed? |
The Neuron cuts through the noise to bring you smart, hype-free takes on the latest AI trends, tools, and breakthroughs. Join 500,000+ professionals from top companies like Microsoft, Apple, Salesforce and more.
|
|
|
Advertise in Daily Tech Insider! Daily Tech Insider is a TechnologyAdvice business
© 2025 TechnologyAdvice, LLC. All rights reserved. TechnologyAdvice, 3343 Perimeter Hill Dr., Suite 215, Nashville, TN 37211, USA. |
|
|
|
