Alex Banks

It's actually quite scary how advanced China's robotics are.

I think Unitree's new robot dog is the ultimate pursuit machine.

Meet the AS2-W:

→ Wheels fused with legs
→ 16kg continuous payload 
→ 30km+ of driving range unloaded
→ It drives on the flat and climbs when it can't
→ Movement trained with RL then transferred to the real world

The robot learned to balance, climb and recover in simulation, then walked out into the real world, crossing rivers and driving down near-vertical cliffs.

We're already seeing robot dogs used in security patrols and military exercises.

This capability is only advancing, especially being able to travel at a max speed of over 6 m/s (over 13 mph), which is insanely fast.

For comparison, the average running speed for an adult typically ranges from 6-8 mph.

But that’s over a flat surface…

Add facial recognition and a weapon to a machine that crosses almost any terrain for 30km without tiring, and you have something that could theoretically hunt a specific person nearly anywhere on Earth.

Last week I wrote about Japan's floating robot companions, soft and faceless, designed to be hugged.

Every robot's form tells you what it's for.

A floating companion is built for connection. A humanoid with hands is built for the home.

But with these dogs, my mind always comes back to Black Mirror's Metalhead.

When I first saw this video, I thought it would be something fun to discuss on a Friday.

Instead it's left me feeling quite uncomfortable for what might be next.

If you liked this post, you'll love the newsletter.

I cover the most important AI developments each week. Subscribe here: https://lnkd.in/eAt_kGef

And follow me Alex Banks for daily AI highlights and insights.

One Claude user consumed $27,000 of compute in 23 days.
They paid $200. Now everyone's limits are getting cut. 
Here’s why this is unsustainable:

Anthropic just finished a two-week promotion doubling Claude's usage limits during off-peak hours.

That promo ended March 27th.

The very next day, they reduced limits during peak hours.

Give with one hand. Take with the other.

Max 20x subscribers paying $200/month are reporting they hit session limits in 3-4 prompts where they'd previously get 20+.

A user on X (@Pranit) called this exact move 11 days before the announcement.

Anthropic’s playbook:

1. Quietly reduce limits
2. Offer a temporary 2x promo that absorbs the transition
3. Let the promo expire
4. The new floor is lower than the old one, but nobody noticed

The reason they can do this is that these consumer AI plans are massively subsidised, and the limits were never defined in the first place.

Whilst the API is totally transparent: $5/million input tokens, $25/million output tokens.

Consumer plans just say "5x more usage" or "20x more usage."

More than what? They've never assigned a number.

Undefined limits mean unlimited flexibility to move the ceiling without anyone pointing to what changed.

Now here's why this is happening.

After Anthropic refused to remove AI safeguards for the U.S. Department of Defence, Claude became the no.1 free app on the App Store.

Over a million new users were signing up per day. That's a great problem to have, until your infrastructure can't keep up.

One power user (@jumperz on X) consumed 1.1 billion tokens in 23 days.

That’s roughly $27,000 in API-equivalent compute on a $200/month plan (135x multiplier).

Insane.

Anthropic burns 70 cents of every dollar it brings in. Inference costs came in 23% higher than their own projections. Breakeven isn't expected until 2027-2028.

This is the Uber playbook.

Subsidise rides until everyone deletes their cab app, then raise prices once dependency is locked in.

Anthropic is doing the same thing with AI compute:

→ Burn through venture capital to buy market share
→ Build workflow dependency
→ Slowly correct the subsidy as users get too embedded to leave

The weekly rate limits that arrived in August 2025 were the first signal. This is the second.

And the competitive pressure is only increasing. Paying subscribers are openly comparing Claude's tight limits to OpenAI, where some plans offer hundreds of requests without hitting a ceiling.

The day Anthropic announced its caps, OpenAI reset Codex usage limits across all plans.

I use Claude every single day. I think it's the best LLM available right now.

But if you're building your workflow around any AI tool, understand what you're actually paying for.

The floor can shift at any time and you won't see it coming.

Follow me Alex Banks for daily AI highlights and insights.

This idea was from my recent newsletter.

Read it here: https://lnkd.in/eyGqxMNy

Anthropic just signed a multi-billion-dollar compute deal with Elon Musk.

Three months ago, he called them "misanthropic".

Now Anthropic gets ALL the compute capacity at SpaceX's Colossus 1 supercomputer (which absorbed xAI earlier this year).

Here's the deal:

→ Over 220,000 NVIDIA GPUs (H100s, H200s, GB200s)
→ More than 300 megawatts of capacity
→ Coming online within the month
→ Directly improves capacity for Claude Pro and Claude Max

Quick primer on training vs inference:

→ Training = teaching the model. Big, periodic compute spend that produces the next version of Claude.
→ Inference = running the model. Continuous compute that serves users and generates revenue.

The deal is for inference, not training.

Anthropic isn't building the next Claude on Colossus 1.

They're using it to serve existing models to far more people, far faster.

Which is why the user-facing benefits hit immediately:

→ Claude Code's 5-hour rate limits doubled (Pro, Max, Team, Enterprise)
→ Peak hour restrictions removed on Claude Code (Pro and Max)
→ Significantly higher API rate limits for Opus models

It also explains why Musk was comfortable handing over the keys.

SpaceXAI already moved its own training workloads to Colossus 2.

On the unit economics, I want to highlight Jamin Ball’s (Partner at Altimeter) napkin math:

Assuming a mix of H100/H200/GB200s at standard rental rates, Colossus 1 generates ~$5B/year of revenue for SpaceXAI.

Now apply Dario Amodei's framework from the Dwarkesh podcast: a 50/50 split between training and inference compute spend, with 60-70% gross margin on inference.

Anthropic could turn that $5B compute spend into ~$15B of revenue. A big win-win.

Anthropic also expressed interest in partnering with SpaceX to develop "multiple gigawatts of orbital AI compute capacity."

In February, Elon was posting that Anthropic "hates Western Civilization." Last week he spent time with their senior team and called them "highly competent."

The demand for Anthropic's models has run so far ahead of supply that renting from a direct competitor is the most rational move available.

They've now stacked compute deals with Amazon (5GW), Google (5GW), Microsoft/NVIDIA ($30B), Fluidstack ($50B), and now SpaceX.

Every one of those deals is solving the same constraints of power, land, and cooling.

Data centres in space are next.

Follow me Alex Banks for daily AI highlights and insights.

I’ll be doing a full breakdown of this deal in my newsletter this Sunday.

Subscribe here: https://lnkd.in/ePSZP6KF

Water usage in AI data centres has been a huge worry for years.

The reality: they use 0.2% of America's daily water, and NVIDIA just pushed that lower.

The reason is a shift in how these sites are cooled.

Traditional cooling towers shed heat by letting water evaporate, so they drink continuously.

NVIDIA puts that at roughly 2.6 million gallons per megawatt every year.

A mid-sized 50MW site gets through about 130 million gallons annually just to stay cool.

Its new Rubin servers change the physics for the better.

They're the first to be 100% liquid-cooled, no fans anywhere, with a coolant that's 75% water and 25% propylene glycol piped straight onto every chip.

→ Coolant enters the chip at 45°C, hotter than a hot tub
→ Leaves at around 55°C, with no drop in performance
→ Already warm, so the heat vents straight outside, no evaporation

The loop is filled once and recirculated for the life of the building.

NVIDIA's figures for a single 50MW site:

→ Over $4 million saved a year on cooling energy and water
→ Cooling can eat 40% of a data centre's electricity, and that drops sharply
→ Six rack units of kit now fit in two, and the 85-decibel fan roar disappears

Two things this doesn't fix that I think are worth calling out.

1. Near-zero water only holds where the climate cooperates.

A site in the Scottish Highlands can reject heat into cool air all year.

The same site in Phoenix still fires up chillers through summer.

NVIDIA's own target is zero water "outside of maybe 1% of the year".

2. This cuts the water used on-site, not the water burned at the power plant feeding it.

A data centre running on wind or solar has an indirect footprint near zero.

One running on coal stays thirsty no matter how clean the loop is.

Fix the cooling and you've solved roughly a third of the problem. The energy source is the rest.

The part I find most interesting is the second-order effects. That captured heat can be piped to warm nearby homes and offices, turning a data centre from an energy drain into a grid asset.

For as long as AI has been mainstream, the story has been AI vs. the planet. The engineering is starting to suggest it doesn't have to be a trade-off.

Follow me Alex Banks for daily AI highlights and insights.

I cover the developments like this that actually matter each week in The Signal.

Subscribe here: https://lnkd.in/ePSZP6KF

This should be illegal.

We’re entering a world where nothing can be trusted online:

I recently came across this video using Kling Motion Control.

It takes your movements and puts them in anyone's body.

Here's how it works:

→ Record yourself doing any movement
→ Use AI to generate a character image
→ Kling 2.6 merges the two seamlessly
→ Your moves, their face

We've now hit the threshold where it's impossible to discern whether someone is human in the digital world.

My takeaways:

The implications are huge in Hollywood:

→ Using someone's likeness without them present
→ Character swapping cost trends to near-zero
→ Reshooting scenes without actors on set

I also see new markets emerging:

→ Individuals renting out their identity
→ Licensing your likeness for content creation
→ Actors selling "performance rights" to their digital twin

Finally, proof of authenticity will become essential infrastructure, not just a nice-to-have.

Sam Altman is already building a global identity verification system using iris scans to fight against fraud and bots.

I believe we'll see identity become the next great asset class.

Follow me Alex Banks for daily AI highlights and insights.

P.S. If you liked this post, you'll love the newsletter.

I help you learn AI each week.
↳ Subscribe here: https://lnkd.in/ePSZP6KF

Video credit: ederxavier3d on Instagram

Your AI company is just a wrapper.

But that's the whole point.

When Chinese company Manus AI released their autonomous agent, one comment kept appearing:

"It's just a Claude wrapper."

The comment was absolutely right. Manus is only a Claude wrapper.

As Peak, a Co-Founder of Manus, clearly stated:

"We use Claude and different Qwen-finetunes" alongside various tools and open-source packages.

Yet somehow this becomes a "gotcha" that diminishes what's been built.

For some context, let's extend this logic:

• Salesforce is an Oracle database wrapper valued at $320 billion
• Stripe is a Mastercard wrapper valued at $70 billion
• AWS is an HPC primitives wrapper valued at $3 trillion

It turns out that everything in tech is a wrapper.

But in AI things get interesting:

→ We thought the application layer would be commoditised
→ We believed value would concentrate in foundation models
→ The reality is playing out differently

The models themselves are rapidly becoming commodities.

As the cost of intelligence trends toward zero, the real opportunity lies in:

1. Infusing unique insights
2. Leveraging proprietary data sets
3. Creating magical user experiences
4. Building network effects and a community
5. Combining the best LLMs with specialised tools

These are the areas where I see a sustainable advantage.

So, yes, your AI company might be "just a wrapper."

But that's exactly where the value is being created.

Follow me Alex Banks for daily AI highlights and insights.

I cover the most important AI developments each week in my newsletter.

You can read the latest issue here: https://lnkd.in/eAt_kGef

Claude just took # 1 and # 3 on the weekly LLM leaderboard.

ChatGPT isn't even in the top 10.

Here's OpenRouter's top 5:

1. Claude Sonnet 4.6: 1.38T tokens (+19%)
2. DeepSeek V3.2: 1.28T tokens (+1%)
3. Claude Opus 4.6: 1.22T tokens (+2%)
4. MiMo-V2-Pro: 1.15T tokens (+90%)
5. Gemini 3 Flash Preview: 1.14T tokens (+8%)

GPT-5.4 sits at # 11 with 564B tokens.

For context, OpenRouter is a unified API that routes developers to 300+ AI models.

It processes trillions of tokens per week, making it one of the best proxies for real-world model preference.

Now pair this with Similarweb's Gen AI traffic share.

12-month shift:

→ ChatGPT: 77.43% → 56.72%
→ Gemini: 6.00% → 25.46%
→ Claude: 1.40% → 6.02%
→ Grok: 7.03% → 3.44%
→ DeepSeek: 3.73% → 3.74%

Claude more than quadrupled its consumer share in 12 months.

Similarweb only tracks website visits though.

API usage and integrations don't show up anywhere in that data.

That's where OpenRouter becomes essential.

Currently ~70-85% of Anthropic’s revenue is generated through API usage.

For example, Claude Code’s run-rate grew from $500M in mid-2025 to over $2.5 billion today.

The products you use every day are increasingly built on Claude: Cursor, Lovable, Manus, and countless others all route through the Anthropic API.

Consumer traffic tells you who's winning the app war.
OpenRouter tells you who's powering the software economy.

Right now, both charts are trending the same direction.

Anthropic continues to win across both fronts.

Follow me Alex Banks for daily AI highlights and insights.

I cover the most important AI developments each week in my newsletter.

Subscribe here: https://lnkd.in/ePSZP6KF

Dario Amodei warned AI would eliminate half of entry-level white-collar jobs.

This week, he said those jobs will multiply, not vanish.

Onstage with JPMorgan CEO Jamie Dimon on Monday, the Anthropic CEO shifted his story.

His argument:

→ Automate 90% of a job, the remaining 10% expands to fill the role
→ Workers become 10x more productive
→ Cheaper services drive more demand
→ The economy creates more work, not less

Economists call this the Jevons Paradox.

Steam engines got more efficient in the 1800s and coal use went up, not down.

Cheaper means more demand, not less.

Compare that to his Axios interview last May, where he predicted half of entry-level white-collar jobs would disappear and unemployment would hit 10-20% within five years.

In "The Adolescence of Technology" (January 2026), Dario actually cited the Jevons Paradox first.

Then he wrote: "It's possible things will go roughly the same way with AI, but I would bet pretty strongly against it."

He listed four reasons AI breaks the pattern:

→ Speed (markets can't adapt fast enough)
→ Cognitive breadth (no neighbouring jobs to switch into)
→ Slicing by cognitive ability (lower-skilled workers get squeezed first)
→ Self-improvement (AI fills its own gaps before humans can)

None of those have changed.

Jevons Paradox took decades to play out for steam engines and ATMs.

Dario himself said AI doesn't have it.

The part that I think is worth taking seriously:

The first-year associate doesn't disappear. But their tedious document review work does.

What's left is the 10% that actually requires human judgment, and there's more of that to do.

The real question now sits on whether people can move up the value chain fast enough.

That's the bit we all need to pay attention to.

Follow me Alex Banks for daily AI highlights and insights.

I cover the most important AI developments each week in The Signal.

Subscribe here: https://lnkd.in/ePSZP6KF

Image source: Fabrice Coffrini/Agence France-Presse/Getty Images

Anthropic said no to the Pentagon.
OpenAI signed the deal hours later.
Here's what happened:

Anthropic drew two red lines with the Department of War:

1. No mass domestic surveillance
↳ AI can now analyse bulk data the government buys on Americans at scale
↳ Currently legal, but only because the law hasn't caught up

2. No fully autonomous weapons
↳ AI isn't reliable enough to automate selecting and engaging targets
↳ No oversight framework exists for removing humans from the loop

For context:

→ Anthropic was the first AI company on the classified cloud
→ Deployed across intelligence, cyber ops, and combat support
→ Forfeited hundreds of millions cutting off CCP-linked firms
→ These two red lines represent ~1% of use cases

The Pentagon gave Anthropic a 3-day ultimatum.

Anthropic refused.

The response:

→ President Trump: "Their selfishness is putting AMERICAN LIVES at risk"
→ Secretary of War Pete Hegseth: Anthropic is a "supply chain risk"
→ Trump ordered every federal agency to cease all use of Anthropic's technology

Then OpenAI entered.

On Friday morning, Altman publicly backed Anthropic's red lines.

By Friday evening, he'd signed the Pentagon deal himself:

→ OpenAI claims the same red lines as Anthropic, plus a third
→ OpenAI asked the Pentagon to offer identical terms to all AI labs
→ Multiple OpenAI employees signed an open letter supporting Anthropic
→ Altman admitted the deal was "definitely rushed" and "the optics don't look good"

The consumer backlash was immediate:

→ "Cancel ChatGPT" went viral across Reddit and X
→ Claude hit # 1 on the App Store, overtaking ChatGPT
→ Claude is now # 1 in Germany, Canada, and other markets

My takeaway:

Not only have I seen my workflows fully transition from ChatGPT to Claude over the last year, but the consumer mass-market is catching up.

I think it’s also important to point out that correlation doesn’t necessarily mean causation here.

Claude didn’t just hit # 1 in the app store as a result of ChatGPT accepting the DOW deal.

Anthropic has simply built a better product where the timing of adoption happened to coincide with this viral uproar of the company denying use of their products for the surveillance of Americans and operation of autonomous weapons.

Ethical positioning now has direct commercial consequences.

Values are becoming a competitive moat.

I covered this idea yesterday in my newsletter.

Be the first to receive it by subscribing today.

Read it here: https://lnkd.in/e898z7SG

From July 12 we’ll see the rise of metered intelligence.

Rather than Fable 5 being included within your monthly Claude subscription, it will instead run under monitored usage limits.

This means you’ll be able to use Fable 5 as much as you like, but you’ll be paying for every single question you ask it.

As a result, a huge emphasis will be placed on the purchasing power of the consumer.

Those who can afford to use it as much as they like will have an “intelligence edge” over just using the likes of Opus 4.8 under their subscription plan.

After using Fable 5 myself consistently since July 1, it’s not as scary as you think. We’re a way off AGI.

Sure, it’s slightly better than Opus 4.8 on long-horizon knowledge work tasks that require significant depth of thought, but the capability gap isn’t yet that high.

However, as new, increasingly intelligent models continue to roll out (think Fable 6,7, etc.), this will grow, and the gap between the capability of those on a basic subscription and those paying by the prompt will grow accordingly.

Will a two-tier society emerge?

Or will the cost of intelligence continue to plummet due to competition and open-source pressures?

Maybe we’ll land somewhere in the middle.

Let’s wait and see.

This should be illegal.

We’re entering a world where nothing can be trusted online:

I recently came across this video using Kling Motion Control.

It takes your movements and puts them in anyone's body.

Here's how it works:

→ Record yourself doing any movement
→ Use AI to generate a character image
→ Kling 2.6 merges the two seamlessly
→ Your moves, their face

We've now hit the threshold where it's impossible to discern whether someone is human in the digital world.

My takeaways:

The implications are huge in Hollywood:

→ Using someone's likeness without them present
→ Character swapping cost trends to near-zero
→ Reshooting scenes without actors on set

I also see new markets emerging:

→ Individuals renting out their identity
→ Licensing your likeness for content creation
→ Actors selling "performance rights" to their digital twin

Finally, proof of authenticity will become essential infrastructure, not just a nice-to-have.

Sam Altman is already building a global identity verification system using iris scans to fight against fraud and bots.

I believe we'll see identity become the next great asset class.

Follow me Alex Banks for daily AI highlights and insights.

P.S. If you liked this post, you'll love the newsletter.

I help you learn AI simply each week.
↳ Subscribe here: https://lnkd.in/ePSZP6KF

Video credit: ederxavier3d on Instagram

Anthropic just measured which jobs AI is actually replacing.

The gap between theory and reality is massive.

Anthropic published a new research paper using its own Claude usage data to track AI's real-world impact on jobs.

What's new:

They created a metric called "observed exposure" that combines theoretical AI capability with actual professional usage data. The results are eye-opening.

→ Computer & Math: 96% theoretical capability. 32% actual coverage.
→ Office & Admin: 94% theoretical. 42% observed.
→ Legal: 88% theoretical. Just 15% observed.

Capability isn't the bottleneck. Legal constraints, verification requirements, and slow enterprise adoption are what's holding back real-world deployment today.

Most exposed occupations:

→ Computer programmers top the list at 75% task coverage
→ Customer service reps follow at 70%
→ Data entry keyers at 67%

But there’s a certain irony at play that I think is worth pointing out.

Programmers are both the most exposed occupation AND the heaviest adopters of AI.

They're actively building and using the technology that automates their own work.

The workers most at risk overall skew older, female, more educated, and higher-paid, earning 47% more on average than their unexposed counterparts.

Graduate degree holders are nearly 4x more represented in the most exposed group.

Despite all this exposure:

→ No meaningful increase in unemployment for high-risk workers since ChatGPT launched
→ But hiring of 22-25 year olds into exposed roles has dropped roughly 14%
→ No equivalent decline for workers over 25

My takeaway:

It’s interesting to see the “disruption” showing up as a hiring freeze vs sweeping layoffs.

But mainstream media much prefer to print “thousands made redundant” to sensationalise headlines.

I also think it’s important to point out the 30% of workers that have zero AI exposure.

Cooks, bartenders, mechanics, lifeguards. The roles AI can't touch are almost entirely physical.

Having a living measure like this helps track how the gap between AI’s theoretical capability and real-world adoption narrows over time.

That gap is where the next wave of disruption lives.

Follow me Alex Banks for daily AI highlights and insights.

I talked about AI’s impact on jobs first in my newsletter.

You get the most important news + analysis in your inbox every Sunday.

Read it here: https://lnkd.in/ei8r5Xyq

Elon Musk spent a decade promising Mars.

Now SpaceX is building a city on the Moon instead.

Why the Moon wins on iteration speed:

→ Launch to the Moon: Every 10 days
→ Launch to Mars: Every 26 months
→ Trip to the Moon: 2 days
→ Trip to Mars: 6 months
→ Moon city: <10 years
→ Mars city: 20+ years

Mars is still on the roadmap. SpaceX plans to begin Mars efforts in 5-7 years.

But Musk's words: "the overriding priority is securing the future of civilisation and the Moon is faster."

Now layer in the bigger picture.

Last week Musk merged SpaceX with xAI in a $1.25 trillion deal.

It’s now the most valuable private company in history with a potential ~$50 billion IPO on the horizon.

Musk wants to launch AI data centres into orbit, arguing terrestrial power grids can't keep up with AI's energy demands.

SpaceX has already asked regulators for permission to launch 1 million satellites for an "orbital data centre system” up from 9,400 today.

It's also worth noting this is part of a broader consolidation of Musk's empire.

xAI absorbed X (previously Twitter) in March 2025. Tesla invested $2 billion in xAI last week while pivoting hard toward AI and robotics.

Investors are already speculating Tesla could eventually fold into the group too.

The Moon is just the starting point.

Follow me Alex Banks for daily AI highlights and insights.

I cover the most important AI developments like this each week in my newsletter.

Subscribe here: https://lnkd.in/ePSZP6KF

Claude just solved the biggest problem with AI.

Memory is now available to Pro and Max users.

What's new:

→ No more repeating yourself every chat
→ Each project has separate memory spaces
→ Persistent context across all conversations
→ Incognito mode for conversations you don't want saved
→ Previously only available to Team/Enterprise customers
→ Claude remembers your projects, preferences, and work patterns

To get started with memory:

1. Go to Settings
2. Navigate to Capabilities
3. Look under Memory section
4. Toggle on “Search and reference chats”
5. Toggle on “Generate memory from chat history”

Bonus: Click on “Memory from your chats” to update/remove memories

Then I recommend asking Claude “What did we work on last week?”

Useful prompts to try with memory:

• "What patterns do you see in my work from our past conversations?"
• "Make unique connections between the ideas we've discussed"
• "Highlight non-obvious insights I might have missed"

Why this matters:

Pair Memory with Claude's new desktop app and you get:

→ Desktop: Always accessible (double-tap access, screenshots, voice)
→ Memory: Always contextual (picks up where you left off)

This turns Claude from a stateless chatbot into a persistent working partner.

My takeaway:

Memory has been the missing link with LLMs.

Other AIs force you to rebuild context in every conversation.

Claude now learns from every interaction and improves with each chat.

This is the difference between a tool you use occasionally and an assistant you work with daily.

Follow me Alex Banks for daily AI highlights and insights.

I cover the most important AI developments each week in my newsletter.

Subscribe here: https://lnkd.in/ePSZP6KF

NEWS: You can now connect Claude and ChatGPT to your business data.

Oracle NetSuite just expanded its AI Connector Service.

Some enterprise platforms are embedding AI as a fixed feature inside their products.

NetSuite is doing both: enabling customers to leverage AI inside and outside the system.

They call their strategy "AI your way."

Here's what they're building:

1. AI Connector Service (built on MCP)

→ Connect Claude, ChatGPT, and popular LLMs directly to your business data
→ NetSuite interface sits inside your AI assistant of choice
→ Governed by the same permissions your team already works under
→ Companion gives you 100+ prompt templates mapped to your role

2. NetSuite Next (phased global rollout)

→ Full platform rebuild with AI and agentic workflows at the core
→ Ask Oracle lets you query your business data in natural language
→ No implementation needed: customers can switch to NetSuite Next with the press of a button

3. SuiteAgents

→ Build custom AI agents directly inside the platform
→ Partners and developers can create their own agentic workflows
→ Agents analyse, automate, and take action on your behalf

The idea is flexibility, so businesses can plug in their preferred AI model and use it in the context of their business data.

NetSuite started 28 years ago after a conversation between Evan Goldberg and Larry Ellison about the future of business software.

It became the first cloud company.

Oracle acquired NetSuite in 2016, and today more than 43,000 businesses run on the system.

Financials, operations, and customer data centralised in one place.

Better data makes for better AI. And that’s the foundation NetSuite provides.

NetSuite runs on Oracle Cloud Infrastructure.

Oracle's AI, Oracle's security, Oracle's scale, all powering the system.

My takeaway:

I was at #SuiteConnect London and sat down with

• Evan Goldberg (Founder and EVP of NetSuite)
• Patrick Puck (Group Vice President, AI Strategy, Engineering and Design at NetSuite)
• Ben A. (CFO at NetSuite customer Yoto)

to dig into what all of this actually means.

Most enterprise software vendors are building walled gardens right now.

They want you using their AI, inside their ecosystem, on their terms.

While continuing to build AI capabilities within the system, NetSuite has also taken another approach.

They've made the system open and composable so businesses can use whatever AI tools they're already comfortable with and use them alongside their actual data.

The AI models are good enough now.

Access to real data and the trust to act on it is where the real value gets unlocked.

And that's exactly where NetSuite is focused.

Follow me Alex Banks for daily AI highlights and insights.

Being polite to ChatGPT is making it dumber.

Saying “please” and “thank you” makes your results worse:

Penn State researchers gave ChatGPT 250 questions across maths, science, and history into five tones from "Very Polite" to "Very Rude" and ran them all through GPT-4o.

Every single comparison favoured rudeness. Not one favoured politeness.

Adding "Would you be so kind" to your prompt literally made ChatGPT dumber. Meanwhile, "You poor creature" produced the best results.

Why? LLMs are trained on vast amounts of human text. In that data:

→ Demanding language = high-stakes, precise communication
→ Polite hedging = casual, low-stakes exchanges

When you add "please" and "no rush," you activate casual mode.
When you're blunt and direct, you activate precision mode.

It gets worse over time.

A second study from Carnegie Mellon tested GPT-4o in three personas:

→ Friendly persona: 64% accuracy
→ Default persona: 71% accuracy
→ Adversarial persona: 71% accuracy

The "friendly" model dropped as low as 61.7% across follow-up rounds. The adversarial model never dipped below 69.7%.

When you tell the AI to be nice, it becomes a pushover.

Now both of these studies tested previous-generation models. GPT-4o, Claude 3.5 Sonnet, Gemini 1.5 Flash. Late-2024 models.

We're now in 2026. The frontier has moved significantly.

Developer Daniel Weinshenker recently tested Claude Opus 4.6 on exactly this question. Polite agent, neutral agent, insulted agent all given the same tasks.

For well-specified tasks, tone made zero difference. All three produced identical solutions.

Newer models have been specifically trained to be more robust to tonal variation. GPT-5.2 auto-routes between reasoning modes based on task complexity, not how rudely you asked.

These models are better at ignoring the noise around a prompt and focusing on the actual instruction.

But the underlying principle hasn't changed.

The reason rude prompts outperformed polite ones was never about rudeness itself.

It was about writing clear, direct, imperative prompts.

Directness is your weapon.

I built a master prompt from all three studies that forces any model into precision mode.

No rudeness required, only clarity.

Copy it for your next important prompt: https://lnkd.in/eNGPpaiM

"He who controls the spice controls the universe."

Baron Harkonnen's line describes the AI compute market almost perfectly.

Anthropic wants to lease AI computing capacity from Meta for up to $10 billion over two years, according to the New York Times.

The Claude maker proposed the deal in June and would pay in monthly instalments of roughly $417 million, with early-exit options on both sides.

The talks are early, but the direction is unmistakable.

→ Anthropic pays SpaceX ~$1.25 billion/month for Colossus 1 in Memphis (3yr agreement)
→ Google also rents SpaceX GPUs at ~$920 million/month
→ Google rents out its own chips too: Anthropic has secured over 1 million of its TPUs, and Meta signed a multibillion-dollar TPU deal in February
→ Meta will spend up to $145 billion on infrastructure this year, more than double last year's $72 billion
→ It has hired Dave Brown, a near two-decade AWS veteran, to build the cloud business Zuckerberg called "definitely on the table" in May

In Dune, the foundational truth is that the entire empire runs on the spice supply mined from the desert planet Arrakis to function.

Whoever controls the supply controls everything downstream.

Compute is the spice of the AI race. And it does two things.

It trains the models, and it runs inference, the process of getting answers back from your prompts. You need both to build and operate the leading intelligence on the planet.

I think it perfectly summarises this next phase of the AI saga, where the owners of compute (SpaceX, Google, Amazon, and now Meta) increasingly set the direction of where the frontier models are heading.

Anthropic is the Guild Navigator here. The most capable intelligence in the universe, suspended in a tank of someone else's spice.

And the landlords collect rent whichever model comes out on top. Even better if it's their own.

I wrote about this idea first in my newsletter. You can read it here: https://lnkd.in/eTqaPgt7

And follow me Alex Banks for daily AI highlights and insights.

BREAKING: Synthesia just raised $200M.

Crazy to think AI avatars will start talking back.

I've been partnering with Synthesia since June.

They just closed a Series E at a $4 billion valuation.

Here's why I’m so bullish on AI avatars:

1. The knowledge problem

→ Companies are drowning in documents, wikis, and training materials
→ Yet employees still can't get the right answer when they need it
→ Upskilling has become a continuous, board-level priority
→ Traditional content can't keep up with the pace of change

2. The AI shift

→ Agents can now understand context and hold real conversations
→ They can coach people through complex scenarios
→ They complete actual workflows vs just generating content
→ We've moved from static video to interactive experiences

3. The opportunity

→ Synthesia started with AI video, the most effective way to teach at scale
→ Now they're turning enterprise knowledge into conversational agents
→ Early customers are already seeing higher engagement and faster knowledge transfer
→ This creates a credible path to a billion-dollar revenue platform

The round was led by Google Ventures with NVentures (NVIDIA’s venture capital arm), Accel, Kleiner Perkins, and NEA doubling down.

My takeaway:

Synthesia is now one of the most valuable AI companies in Britain.

This transforms video from a one-way communication into a two-way interactive conversation.

AI will drive the marginal cost of creating content to zero.

We are now generating video through code rather than recording with a physical camera.

Instead of being a single static medium, video can now change and adapt depending on who’s watching.

The window for this opportunity is open now.

Synthesia is positioned to define the category.

Excited to see what Victor Riparbelli and the team build next.

Follow me Alex Banks for daily AI highlights and insights.

You can now give Claude memory with projects.

No more starting from zero every session.

Every Cowork session used to begin the same way.

You’d have to re-upload files, re-explain context, and re-state what you’d already covered.

Projects change this by bundling four things into one workspace:

→ Instructions: rules Claude follows for every task in scope
→ Context: reference docs, transcripts, and folders Claude reads from
→ Scheduled tasks: recurring work that runs automatically in the background
→ Memory: scoped to the project, so nothing leaks across workstreams

The simple test for Skill vs Project:

"Does this work need memory of what came before?"

If no, it's a Skill (a recipe for repeatable tasks).
If yes, it's a Project (a workspace for ongoing work).

Skills are the tools in the workshop. Projects are the workshop itself.

5 steps to set one up:

1. Create a project (start fresh, import from Chat, or use an existing folder)
2. Add instructions defining the goal, terminology, and rules
3. Add context like reference docs, transcripts, and PDFs
4. Run your work inside the project
5. Stack it with Skills for repeatable formatting

This is the layer that finally makes Claude feel like a real partner for thought.

I broke down how to set up Claude Projects in full detail here: https://lnkd.in/e8dz8pRk

Ex-Google CEO was booed while discussing AI in his commencement speech.

Eric Schmidt told the class of 2026 that AI would touch every profession.

The University of Arizona turned on him in real time.

"I know what many of you are feeling about that. I can hear you," he said, before talking about the fear that the future has already been written.

He wasn't alone.

Two other speakers got the same treatment this graduation season:

→ Gloria Caulfield at UCF: AI is the next industrial revolution
→ Scott Borchetta at Middle Tennessee State: AI is rewriting the industry

All three reach for a framing around AI as inevitable progress.

This is something that used to be reliable applause 2-3 years ago.

But to a room of people about to job-hunt in an AI-shaped market, it now lands like a threat wearing the veil of a pep talk.

Pew research found that half of Americans are more concerned than excited about AI, up from 37% in 2021, with just 10% mainly excited.

Among AI experts, that figure is 47%.

We’re not quite at the Luddite riot stage yet, but the backlash is now clearly visible, as the younger workforce feels deeply passionate about its evolving capabilities.

Let’s see how the following five years play out.

Follow me Alex Banks for daily AI highlights and insights.

Want to start leveraging AI?

Subscribe to my newsletter: https://lnkd.in/ePSZP6KF

Video credit: WSJ News on YouTube

NEWS: ElevenLabs just dropped massive AI updates.

Voice was just the starting point.

Platform 1: ElevenAgents

↳ Expressive Mode: their most emotionally intelligent model yet
↳ Turn-taking system that reads emotional cues from how you speak
↳ 70+ languages, live with the Ukrainian Government and US local governments

Platform 2: ElevenCreative

↳ Generate images, video, music, voiceovers, and sound effects from prompts
↳ Flows: node-based video editor for automated content pipelines (coming soon)
↳ Music Marketplace and Finetune: publish tracks, earn royalties, or generate in your style

Platform 3: ElevenAPI

↳ Direct access to all foundational models
↳ Voice, dubbing, transcription, sound effects, speech-to-speech
↳ Build whatever you need

I spoke with Carles Reina (4th employee and first GTM hire) this week.

"Talking to technology is a lot more natural and engaging and quicker and easier than actually writing to technology."

What he said was spot on. Voice is how humans were built to communicate.

Everyone's obsessing over which model is smartest.

ElevenLabs is focused on something else entirely.

Own the voice and interface layer and you're the infrastructure everyone relies on.

I personally now speak to AI models far more than I type.

Voice is overwhelmingly the interface of the future.

Especially as now agents are capable of reading your emotional state and responding with empathy.

The smartest AI means nothing if it still feels like talking to a machine.

ElevenLabs understood that before anyone else.

Follow me Alex Banks for daily AI highlights and insights.

P.S. I did a full breakdown in my newsletter.

Read it here: https://lnkd.in/eK3gBRbD

#ElevenAgentsPartner

This meme is everywhere right now.

And it's asking the wrong question.

It presents two paths.

AI either succeeds and destroys jobs, or fails and crashes the economy.

But both paths are already happening simultaneously.

AI is succeeding AND displacing workers.
AI is falling short AND companies are over-investing.

Citrini Research asked the perfect question earlier this year.

"What if AI bullishness is right, and that's actually bearish?"

It went mega-viral on Substack with over 8,000 likes.

Fear sells better than hope, especially when the technology is this new and this uncertain.

But here's what the meme gets wrong. It assumes a binary outcome.

Much like the Industrial Revolution transformed work rather than abolished it, I think we'll see the same with the rise of AI.

Sure you had the Luddite riots, decades of wage stagnation, massive social upheaval.

Yet new roles were created, old ones faded, and our definition of "work" shifted entirely.

The difference this time is that AI is coming for both physical AND knowledge work simultaneously.

Blue collar and white collar all at once.

The transition will be the hardest part.

It's messy, very uncertain, and requires an awful lot of preparation.

→ Upskilling and retraining at an unprecedented scale
→ Thoughtful leadership that amplifies humans rather than replaces them
→ Rethinking how society distributes resources AND meaning

Sure you can give someone UBI, but you can't give them meaning.

Someone asks "what do you do?" and you answer with your job.

Strip that away, and the question becomes “who am I now?”

That's the path worth figuring out.

Follow me Alex Banks for daily AI highlights and insights.

I explored this idea in depth in my recent newsletter.

Subscribe here: https://lnkd.in/ePSZP6KF

Figure AI just went from 1 robot a day to 1 robot an hour in 120 days.
This week alone, they'll manufacture 55 humanoid robots at their BotQ facility.

The numbers:

→ 24x throughput improvement in under 120 days
→ 350+ Figure 03 robots delivered to date
→ 9,000+ actuators produced across 10+ SKUs
→ 500+ battery packs shipped at 99.3% first-pass yield
→ 80%+ end-of-line yield, improving weekly
→ 150+ workstations on custom manufacturing software

Each robot now passes 80+ functional verification tests.

Thousands of squats, shoulder presses, and jogs to surface early-cycle failures before shipping.

In 2023, at Tesla Investor Day, Elon Musk said: "I think we might exceed a one-to-one ratio of humanoid robots to humans."

In 2024, he predicted humanoid robots would outnumber humans entirely by 2040.

At Davos this January, he doubled down: "There will be more robots than people."

Manual labour is becoming a manufacturing problem.

For 200 years, every economic shift hit the same wall: how fast can we train workers?

Humanoid robots break that constraint entirely.

Workers no longer need to be born, raised, and educated.

They get assembled.

The humanoid form makes the most sense given the world was made by humans for humans.

And every robot rolling off the line generates more training data for the next.

Figure is shipping the labour force of the future.

I'll do a full breakdown of the manufacturing economics in this Sunday's newsletter.

Subscribe to The Signal here: https://lnkd.in/ePSZP6KF

And follow me Alex Banks for daily AI highlights and insights.

Anthropic is the benchmark right now.

Code with Claude in London was incredible.

This was their first developer conference outside the US and their first in Europe.

I use the term “developer” loosely here.

The demos I saw, both on stage and in the booths, show that natural language is becoming the universal language for building ANYTHING with AI.

You just need an idea and the patience to iterate.

This means that, essentially, everyone is now a “developer”.

They dropped two big updates that actually fit incredibly well with the concerns of the European market.

1. Self-hosted sandboxes
↳ Tool execution moves onto the infrastructure you control
↳ Your files, repos and packages never leave your perimeter
↳ Bring your own setup or start with Cloudflare, Daytona, Modal or Vercel

2. MCP tunnels
↳ Agents reach private databases, internal APIs and ticketing systems
↳ A lightweight gateway makes a single outbound connection
↳ This means that nothing is exposed to the public internet

The agent loop stays on Anthropic's side (orchestration, context, error recovery). Only the sensitive work moves to yours.

A great insight I got from speaking with Katelyn Lesse (Head of Engineering, Claude Platform, Anthropic) and Angela Jiang (Head of Product, Claude Platform, Anthropic) was that Europe and London tend to be far more security-conscious than San Francisco.

That’s why it made a ton of sense to release both these features here on the ground.

It hands organisations a framework to ship agents that are secure and satisfy internal benchmarks, and that's exactly what unlocks genuinely meaningful workflows here.

It also tracks with how Anthropic works internally.

The majority of their code is now written by Claude Code, with engineers focused on architecture and orchestration.

This is where we’ll see the next big focus for builders.

Using natural language to instruct agents. The agents do the doing.

Then humans do the thinking and understand which tools, architectures, and frameworks to build with.

It became abundantly clear that there’s never been a better time to build something meaningful than today.

Here’s to the next Code with Claude 👋 #ClaudePartner

Follow me Alex Banks for daily AI highlights and insights.

Stop forcing one AI to do everything.

Here's how I choose the right model for each task.

One of the most common questions I get:

"I've got ChatGPT, Claude, and Gemini, which model should I use for what?"

Here's how I actually think about it:

I treat LLMs like a toolbelt.

My current task → model map:

1. Long-form writing
↳ Default: Claude Opus 4.5
↳ Backup: ChatGPT 5.2 Thinking

2. Deep research
↳ Default: ChatGPT 5.2 Pro
↳ Backup: Gemini 3 Pro

3. Problem solving & complex reasoning
↳ Default: Grok 4.1
↳ Backup: ChatGPT 5.2 Thinking

4. Learning
↳ Default: Gemini 3 Pro + Guided Learning
↳ Backup: ChatGPT 5.2 + Study & Learn

5. Coding
↳ Default: Claude Opus 4.5
↳ Backup: Claude Sonnet 4.5

A few principles I've found useful:

• Task first, model second
• Pairs, not monogamy (I use 2-3 models every day)
• Latency, cost, context > benchmarks
• Always have a default AND a backup

My takeaway:

"Which model is best?" is the wrong question.

The right question: "What's the job I need done?"

Match the tool to the task. Your output quality will 10x.

I did a full breakdown with my default prompts and setups for each job.

Read it here: https://lnkd.in/eShuwmCt

Underneath Wimbledon is a bunker where IBM's AI reads every match live.

The scale is far bigger than I expected.

Last week I spent the day inside it with Fred Baker, who leads IBM's sports work across EMEA, getting a first-hand look at the AI that underpins The Championships.

It’s truly remarkable what’s going on under the surface, and I don’t think many people understand the insane scale of the operation they’re running.

1. It sits on a 36-year relationship where 25 million data points are gathered every single year to produce one of the richest datasets in the world of sport. It was collected to power scoreboards and broadcast stats. Nobody could have known it would one day become fuel for AI.

2. Recent advances in AI and LLMs have made this dataset truly useful. AI was introduced to Wimbledon in 2017 and has since grown into a powerhouse of insights available at the touch of your fingertips. In 2026, Live Likelihood to Win tracks each player's probability to win point by point, the introduction of Key Moments provides more context and explainability, and Match Chat catches you up on any match in seconds, now with photos and clips pulled in.

3. Fred told me about IBM Bob, their AI development accelerator that has been able to map Wimbledon's archive of 15,000+ digital assets (think old photographs from the 1950s) into a knowledge graph that can be tagged, and key metadata can be extracted. This turns a legacy archive into something meaningful and useful today. Work that would normally take 4-5 specialists months is now done by one engineer in 4 weeks.

It reminds me of the Italian term “Sprezzatura”. It’s the idea of a swan paddling frantically beneath the surface, yet above water, it’s as graceful as ever.

This is the exact vibe I got.

They call it Court 19: the bunker where the numbers are crunched and the insights are birthed.

The app, the crowd, and the fan experience exist and are so powerful because of what happens when nobody’s watching.

That’s how AI becomes meaningful.

Complexity is hidden beneath the surface.

On top is an effortless experience that anyone can interact with.

#IBMPartner #Wimbledon #AI

Humanoids are dominating the headlines, but I think AI companions will be a bigger market.

The one from Japan floats above your head like a small whale.

Mingyang Xu and his team at Keio University's Graduate School of Media Design have spent three years building soft floating robots.

Their prototype uses a helium envelope and two flapping fins instead of propellers.

→ 70 grams total weight
→ 0.78m wingspan, narrowing to 0.45m 
→ Fits through a standard doorframe
→ Fins flap between 0.3 and 1.5Hz, close to silent
→ No rotors, rigid edges or face

In their study, 24 people met it for the first time and nobody was told to touch it.

22 of them did anyway.

They patted it, stroked it, cradled it, hugged it, and pressed their cheeks against it.

20 of the 24 reported feeling calm afterwards. One participant told the researchers that if a drone broke they would just buy another, but if this broke they would feel heartbroken.

I think this highlights the bigger “problem to be solved”.

It’s not just the surface-level tasks of doing the dishes, cleaning, or tidying up the home that humanoids will likely take over.

It’s the case for companionship.

Just take Groove X, which has sold Lovot since 2019 at roughly $3,000.

It deliberately performs no household tasks whatsoever, and in one observational study, Lovot owners showed higher baseline oxytocin than non-owners.

The WHO puts loneliness at 1 in 6 people worldwide and links it to roughly 871,000 deaths a year. In Japan, single-person households hit 44.3% by 2050.

Humans have had a free choice of companion for thousands of years, and we picked animals over something shaped like us.

Don’t get me wrong, both humanoids and robotic companions serve a distinct purpose.

But I think many people are sleeping on the market opportunity for the former.

At the end of the day, nobody buys a dog to do the dishes.

Figure's robot just taught itself to move like a human.

Robots will be living with us sooner than you think.

Their new AI system "Helix 02" controls Figure 03's entire body as one continuous behaviour.

Walking. Balancing. Manipulating. All on a single neural network.

What makes this a leap forward:

→ Trained on 1,000+ hours of human motion data
→ Palm cameras and fingertip sensors can feel objects as light as a paperclip
→ 4 minutes of autonomous dishwasher loading with zero resets or human intervention

The human-like details that caught my attention:

→ Uses its hip to shut a kitchen drawer
→ Kicks the dishwasher door up with its foot
→ Selects the wash program and starts the cycle

Six months ago, Figure 02 was only moving its upper body.

Now Figure 03 walks, balances, and manipulates as one fluid behaviour.

This is a serious step up from factory parcel sorting to generalised domestic capability.

As founder Brett Adcock stated, this has been a year-long effort to re-align their AI stack for long time horizons and complex manipulation.

Now that robots can handle delicate tasks like extracting pills from a medicine box and dispensing precise liquid volumes, we're entering a new phase of home robotics.

I personally can't wait to get my hands on one.

I did a full breakdown of Helix 02 in my latest newsletter.

Get practical AI workflows and tutorials for busy professionals + weekly news analysis.

Read it here: https://lnkd.in/eSaQhyQd

Meta just acquired Manus.

Zuckerberg wants to win. Badly.

When Manus launched in March 2025 the internet dismissed it as "just a Claude wrapper."

Now Zuckerberg is paying ~$2 billion to own it:

→ 147 trillion tokens processed
→ 80 million virtual computers created
→ State-of-the-art performance on real-world AI benchmarks
→ All achieved in just a few months

What makes Manus different:

• Self-directed operation without waiting for instructions
• Multi-agent architecture with specialised sub-agents
• End-to-end task execution from research to deployment

Manus will continue operating its subscription service while integrating directly into Meta's AI products.

The goal is to bring autonomous agents to billions of users and millions of businesses.

This fits Zuckerberg's pattern perfectly:

→ $14.3B for 49% of Scale AI
→ $200M package for Apple's former AI Chief
→ $100M+ offers to poach from OpenAI, DeepMind, Anthropic

People will work for Zuck if the price is right.

My takeaway:

Everyone said the value would concentrate in foundation models.

The reality is playing out differently.

Turns out the application layer is where the real value lives.

Zuck himself highlighted: "The rest of this decade seems likely to be the decisive period."

He's not waiting around.

I'll be doing a full breakdown of this acquisition in my newsletter this week.

Subscribe here: https://lnkd.in/ePSZP6KF

And follow me Alex Banks for daily AI highlights and insights.

Sam Altman said ads were a "last resort."

That day has arrived.

OpenAI is introducing advertising to ChatGPT's free and lower-paid tiers.

Here's what they're promising:

→ Ads won't influence ChatGPT's responses
→ Conversations stay private from advertisers
→ Premium tiers remain ad-free (for now)

The OpenAI timeline:

→ March 2025: Raised $40B at $300B valuation
→ December talks: New $100B round at $750B valuation
→ Today: Burning through cash at an extraordinary rate

Subscription revenue alone isn't cutting it.

This is the fundamental law of the web: If the user doesn't pay the bill, the advertiser does.

Back in May 2024, Sam sat down for a fireside chat at Harvard University.

When asked about advertising, he remarked: "I kind of think of ads as like a last resort for us as a business model."

14 months later, here we are.

No one could have predicted how capital-intensive this race would be.

Add in fierce competition and there's little pricing power left.

My takeaway:

I'm not sure ads alone can cover their losses.

Performing well in advertising often pushes companies toward aggressive data collection.

This is something Google is often criticised for.

Eventually, compute will get cheaper.

But right now it's all about aggressive build-out at any cost.

Once ad revenue becomes material to the business, the incentives shift.

That's just how it works.

The LLM market is competitive enough now that friction like this could accelerate the shift to Claude, Gemini, or Grok.

OpenAI built the most used AI product in history.

Monetising it without eroding trust is the real test.

Would ads make you switch?

I did a full breakdown on this in yesterday's newsletter.

Read it here: https://lnkd.in/en4Mu7ei

Ford hired back 300 human engineers after its AI failed to deliver.

The company just topped the US quality charts for the first time since 2010.

Ford is now the No. 1 mainstream brand in JD Power's 2026 Initial Quality Study.

→ 15th place in 2023 to 1st in 2026
→ 152 problems per 100 vehicles, beating Nissan (156) and Buick (162)
→ 41 fewer problems than last year, the biggest jump of any mass market brand
→ F-150, Mustang and Super Duty each won their segment

In December 2024, Ford rolled out 900 AI-powered cameras across its plants to catch defects on the line.

CEO Jim Farley predicted AI would "leave a lot of white-collar people behind."

To handle quality control, Ford trimmed its workforce and fed the machines enough design data to catch defects before cars touched the floor.

COO Kumar Galhotra admitted Ford leaned on automated systems with disappointing results.

VP of Engineering Charles Poon also said, “Mistakenly, we thought that by just introducing artificial intelligence and ingesting the design requirements that we had, that would produce a high-quality product.”

So Ford brought back its "graybeard" engineers.

Some from retirement, others from suppliers.

Their new job was to run design reviews and mentor younger staff, while retraining the AI that had fallen short.

The veterans held something the training data never captured.

That is: clear judgement.

After enough product cycles, you stop reading the spec and start noticing the thing that feels wrong before you can explain why.

That discernment is where AI keeps hitting a wall.

Now don’t get me wrong, machines handle scale and pattern brilliantly.

But an AI system can only judge based on what it has already been shown. It never built its own sense of what good looks like.

The real value lies in the judgement that determines what the machine should be looking for in the first place.

Pair that human discernment with AI's speed and reach, and you get the combination everyone overlooked in the first place.

Follow me Alex Banks for daily AI highlights and insights.

Want to stay ahead of the curve in AI? Subscribe to The Signal: https://lnkd.in/ePSZP6KF