- Everyday AI
- Posts
- Ep 740: Everything Is Fake: How Your Company Can Leverage Human Expertise and Fight AI Workslop (Start Here Series Ep 15)
Ep 740: Everything Is Fake: How Your Company Can Leverage Human Expertise and Fight AI Workslop (Start Here Series Ep 15)
Anthropic Lets Claude Control Your Computer, Apple Teases AI at WWDC, Google DeepMind Brings AI to Robots, and more.
Everyday AI
March 24, 2026
š Subscribe Here | š£ Hire Us To Speak | š¤ Partner with Us | š¤ Grow with GenAI
Outsmart The Future
Today in Everyday AI
8 minute read
š Daily Podcast Episode: Everything online is starting to feel fake. In Episode 15 of our Start Here series, we break down how to use AI without losing trust. Give todayās show a watch/read/listen to learn more.
šµļøāāļø Fresh Finds: Anthropic Claude Agents Tackle Complex Physics, Google Expands Drive File Support, OpenAI Launches File Library Feature, and more. Read on for Fresh Finds.
š Byte Sized Daily AI News: Anthropic Lets Claude Control Your Computer, Apple Teases AI at WWDC, Google DeepMind Brings AI to Robots, and more. Read on for Byte Sized News.
šŖ Leverage AI: AI is scaling your content, but it might be killing your credibility. Hereās how to regain trust. Keep reading for that!
ā©ļø Donāt miss out: Miss our last newsletter? We covered: OpenAI Seeks New Investors, Meta Builds AI for Leadership, OpenAI Doubles Down on Enterprise Growth, and more. Check it here!
Ep 740: Everything Is Fake: How Your Company Can Leverage Human Expertise and Fight AI Workslop (Start Here Series Ep 15)
You're polluting the world with AI Workslop and you don't even know it. šļø
In a world were everything is free and fake -- or, AI -- it's easy to just throw unlimited spaghetti at the wall and see what sticks.
But there's a downside in just blindly rubber stamping those generic outputs from LLMs. And it's worse than the workslop epidemic. It's losing trust.
So, how can your company survive and thrive in an AI world where everything is fake?
Also on the pod today:
ā¢ 90% of content soon AI? š¤
ā¢ Trust crisis: consumers losing faith š¬
ā¢ Human vs. AI: can't tell! šµļø
Itāll be worth your 33 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Hereās our favorite AI finds from across the web:
New AI Tool Spotlight ā Drift is Claude Code for Robotics Simulations, Maestri is An infinite canvas where your coding agents work in concert, NextPhone Picks up your calls, answers questions, and books directly into your calendar.
Long Running Claude ā Agents rebuilt a differentiable Boltzmann solver to sub-percent accuracy in days. Curious?
Noota Launches Talent Platform ā Noota Talent automates intake-to-offer with five agents, producing a scored, screened shortlist in under 24 hours. Click to Learn More
Figma MCP Tool ā Figma is launching an open beta that lets AI agents design directly on the Figma canvas, using a new use_figma MCP tool and teachable skills.
OpenAI Library ā OpenAI now auto-saves your uploaded and generated files to a Library you can reuse in chats. Want to know how to add, delete, or download them?
Gemini Calendar Tools ā Gemini Apps can manage events across multiple phone calendars, with some limits on invites and details.
Google Drive ā Google Drive now supports more file types for attachments, so sharing across teams and apps is easier.
PlayerZero Launch ā PlayerZero builds autonomous "AI production engineers" that trace failures to the exact line, generate fixes, and push one-tap approvals, collapsing detect-to-fix into minutes.
GAP Using Gemini ā Gap is now the first major fashion retailer to let customers complete purchases inside Googleās Gemini, using AI-powered shopping experiences.
Anthropic Teases Phone Use ā Anthropic appears close to letting Claude control phones, so assistants could make calls and interact with mobile apps.
AI Scam Lookout ā Scammers are using AI to create deepfake voices and texts, and tech companies including OpenAI are racing to build tools to detect and block them. Curious which defenses are working and what to watch for next?
Cursor Instant Grep ā Cursor claims its AI code search is 99% faster than traditional grep, promising instant, context-aware results for developers. Curious how it actually speeds up real-world code hunts?
1. Anthropic lets Claude control your computer š„ļø
Anthropic announced this week that Claude can now execute tasks on a userās computerāopening apps, navigating browsers and filling spreadsheetsāafter being messaged from a phone via its Dispatch feature.
The move follows the industry push toward always-available, agentic AIs driven by the viral OpenClaw trend and comes as competitors and chipmakers race to ship similar tools. Anthropic warns the capability is still early, will ask permission before accessing new apps, and includes safeguards while acknowledging risks and potential mistakes.
2. Apple to Spotlight AI at WWDC, June 8ā12 š”
Apple confirmed its Worldwide Developers Conference will run June 8ā12 both online and at its Cupertino campus, and the big takeaway is a clear shift toward AI, with the company promising updates across iOS, macOS, watchOS and tvOS.
The timing matters: Apple has been quietly building an AI stack including its Foundation Model framework and recent partnerships with Googleās Gemini and thirdāparty coding agents, so WWDC looks set to reveal a more capable Siri and deeper onādevice and cloud AI features.
3. Agile Robots and Google DeepMind link up to put Gemini models into factory robots š¤
Agile Robots and Google DeepMind announced a strategic research partnership today to integrate DeepMindās Gemini Robotics foundation models with Agile Robotsā industrial hardware, aiming to accelerate adaptable, reasoning robots for manufacturing and other high-value industrial uses.
The collaboration will combine real-world deployments, data collection and joint model training to create a scalable feedback loop where operations improve models and better models expand automation.
4. Luma debuts Uni-1: one model for both image creation and precision edits āŗļø
Luma today rolled out Uni-1, a unified image model that lets users either generate brand-new images or surgically edit existing ones with precise control.
The system distinguishes Create versus Modify modes, supports up to nine reference images with explicit role labels, and adds seed-driven reproducibility plus nine aspect ratios for consistent outputs.
5. Alibaba Unveils New AI Chip for Agents š»
Alibaba today revealed a purpose-built AI chip designed to power agent-style applications, marking a timely push as cloud and AI vendors race to deliver hardware optimized for autonomous assistants.
The announcement emphasizes lower latency and improved efficiency for running large models and multi-step agent workflows, positioning the chip as infrastructure aimed at production-ready AI services rather than research prototypes.
6. Meta snaps up Dreamer team to boost AI agents š
According to Bloomberg, Meta has hired the founders and core team of startup Dreamer, including Hugo Barra and David Singleton, to join Meta Superintelligence Labs and work on AI agents, with Dreamer remaining a separate legal entity and licensing its technology non-exclusively.
The move is timely as Meta doubles down on agents as a strategic priority, following recent buys like Manus and Moltbook, and signals continued aggressive spending on AI talent and infrastructure.
6. Microsoft and NVIDIA unveil AI toolkit to speed nuclear energy buildout ā”
Microsoft and NVIDIA announced a collaborative AI platform aimed at collapsing the time and paperwork that hobble nuclear projects, promising tools that streamline permitting, accelerate design with digital twins, and optimize AI operations on Azure.
The move is timed to address a surge in electricity demand and supply-chain reindustrialization that require rapid, scalable, and carbon-free capacity.
Your customers donāt trust you anymore. Yikes.
A recent Salesforce survey reveals 72% of consumers trust companies less than they did a year ago.
The actual reason your buyers are bailing?
Your team is flooding the market with automated AI work slop. (And sorryā¦. your leadership is prolly cheering it on.)
The brands dominating your industry right now are not just buying software seats for AI. They are fundamentally rebuilding how their smartest people interact with the machine.
Most companies treat large language models like magical vending machines. They feed generic prompts into the slot and straight up hope a polished corporate strategy pops out. Instead, they just get a regurgitated pile of bland text that makes everyone go, āgood enough.ā
We broke this down on today's Everyday AI Start Here Series and revealed exactly how to replace passive human-in-the-loop failures with active expert-driven workflows.
More human expertise. Less AI slop. Regain the trust. Letās get it.
1. Stop The Rubber Stamp š
A massive 72% of business leaders just flagged AI-enabled fraud as a top operational challenge in a new Experian forecast.
(Yeah, we know. Another 72% stat back to back. lolz)
But the real internal fraud is happening right under your nose.
We are watching enterprises jam automation into workflows that were already held together with duct tape and prayers. The typical setup involves a technical champion building a workflow, while the actual domain expert just passively rubber-stamps the final output.
This massive disconnect guarantees your deliverables will sound like a robot wrote them while sleepwalking. You cannot expect a generic human-in-the-loop checkpoint to catch the nuances of your proprietary business logic.
Instead, you gotta stop treating your smartest employees like automated toll booth operators. The winners in this space are forcing their domain experts to sit down and literally shape the model's chain of thought before a single draft gets generated.
If you are just checking the final output, your competitive moat is already cooked.
Try This
Pull your three most recent AI-generated client proposals and sit down with your top sales director.
Force them to ruthlessly flag every single sentence that sounds generic, unverifiable, or kinda sus. When they tell you how they would have naturally pitched it, map that exact reasoning back into your core system instructions.
Do not let your IT admins build these prompts in a vacuum without the people actually closing the deals.
Make this ruthless audit a mandatory monthly routine for all external communications.
2. Build Expert Driven Loops š„
SmythOS data recently proved that AI content with human strategic oversight performs more than four times better than fully automated garbage.
Four times.
That is the new baseline fam. But you ain't gonna hit those numbers by just having someone proofread the final paragraph right before clicking send.
The most AI-native organizations are pouring intense human expertise into the actual assembly line. They do not just document what decisions get made.
Nope.
They meticulously map out exactly why those decisions get made in the first place. And when teams pitch new automated workflows? Leadership straight up forces the subject matter experts to validate the model's reasoning process before deploying a single thing.
If your baseline process is a mess, adding AI is just gonna help you fail at scale. You need to transition from passive checkpoints to active expert-driven loops immediately.
Try This
Grab your most senior domain expert and have them shadow the person building your automated workflows for one afternoon.
Force them to read through the model's chain of thought together, step by step, instead of just marveling at the shiny final output. You will immediately spot where the AI loses its mind and hallucinates.
(And real talkā¦. it prolly goes off the rails entirely by step two.)
Turn this shadowing session into a bi-weekly ritual to keep your models grounded in absolute reality.
3. Audit Your Slop Factory ā”
A recent McKinsey study shows only 6% of companies actually qualify as AI high performers right now.
The other 94%?
They are just mass-producing work slop because the economics of fast and cheap are way too tempting to ignore. We are watching businesses willingly trade their hard-earned reputations for the ability to generate a mediocre blog post in four seconds.
Europol projects that up to 90% of online content might be synthetically generated by the end of this year. When the entire internet becomes a graveyard of uninspired garbage, authentic human expertise becomes your only viable lifeboat.
You gotta relentlessly hunt down any generic content your brand is rubber-stamping and scorch it from the earth.
Real talk here. If you are not actively injecting your company's unique point of view into these systems, your brand is effectively invisible. You know the exact vibe. That generic corporate speak that puts everyone to sleep.
Try This
Run a blind taste test this Monday morning with your executive team using your own marketing materials.
Print out three pieces of your recent AI-generated content alongside one piece written entirely by your top performer. If your leadership team cannot instantly spot the human-crafted asset, your entire production pipeline needs surgery.
You cannot throw another dollar at software until you fix this massive gap.
Do this quarterly so y'all never lose the scent of your own authentic brand voice.
Reply