Goldman Sachs new report "An AI Job Apocalypse?"

Says AI could displace 15mn workers, but not cause mass unemployment.

Estimates AI could lift productivity and GDP by 15% after full adoption.

The clearest pressure sits in customer support, back-office work, claims, billing, and entry-level tasks.

Goldman’s own labor model finds a small current drag of 16K jobs monthly.
Younger workers look more exposed because AI can absorb the basic tasks that teach judgment.

College graduates are not collapsing yet, but their industries adopt AI faster than others.

Joseph Briggs, GS global economists still expects more than 9% of workers to face displacement across 10 years.

MIT professor Daron Acemoglu is more cautious, expecting 2-4% job losses across 5 years. His concern is that current AI replaces routine office tasks more often than it supports workers.

Neil Thompson of MIT adds a missing filter: capability does not mean reliability, access, or cheap deployment. A chatbot can answer broadly, but businesses need accuracy, private data, and repeatable workflows.

The xAI team really cooked with Grok 4.5.

The price to performance ratio is honestly kind of insane. Getting frontier-level performance this close to open source hosted pricing is a huge deal. AI inference just got a lot more accessible. 🔥

Both Elon Musk and Demis Hassabis has said robotics is bottlenecked by hands and robots will not become useful until hands improve massively.

And now 1X Technologies is preparing to reveal a robotic hand that may surpass human finger control.

The big deal here is the controllable force across the small joints that make dexterity possible. A 10-20DoF hand would sit near the hard boundary between compact hardware and humanlike control.

The design appears built around Bowden tubes, which route force through cable-like tendon paths.
Real intelligence may be hiding in the routing, not the fingers.

That architecture lets actuators sit away from the fingers, keeping the hand lighter.

A human hand uses the same broad trick, with muscles pulling tendons across joints.

1X is applying that idea to humanoid robotics, where compact hands must handle real objects.

The visible side reportedly shows 20 tendons, suggesting 10 paired actuator paths.

A matching hidden side could raise the system to 40 tendons and 20 actuators.

The lower estimate points to coupled finger joints, where one tendon path drives several links.

The higher estimate points to independent IP control, where each finger segment can move separately.

That difference separates a strong gripper from a hand that can shape contact precisely.

The catch is calibration, because tendon friction, stretch, and backlash can erase beautiful mechanics.

If 1X has solved that reliably, the hand could become a serious advantage for NEO.

39 gold miners sit in the world's top 100 mining companies.

Just 22 copper miners out earn all of them.

By share of $730B in combined 2025 mining revenue:

1️⃣ Copper: 22 companies, 28.1% of revenue, the single largest share of any commodity

2️⃣ Gold: 39 companies, more producers than any other metal, but only 21.3% of revenue

3️⃣ Iron ore: just 5 companies, still out-earning gold at 22.4%

4️⃣ Glencore alone earned $63B, more than the entire coal sector combined

Volume isn't value. The revenue crown belongs to copper and iron ore, not gold.

Grok 4.5: The Complete Overview

• How to set up (for free)
• Real use cases
• Benchmarks & comparisons
• Tips & more

How to turn Hermes into your Google Workspace assistant (step by step).👇

Connect Hermes to:

- Gmail
- Google Calendar
- Google Drive
- Google Notes
- Google Sheets
- Google Meet

I love money.
Always have.
I think in profit margins, not feelings.
Every decision's an ROI calculation.
If it doesn't pay me, it drains me.
I don't chase passion, i price it.
Money isn't everything.
But it's the scoreboard.

You Made $1,000,000.

Here’s What’s Left After Taxes

🇦🇪 United Arab Emirates -$1,000,000
🇵🇦 Panama - $1,000,000
🇵🇾 Paraguay - $1,000,000
🇨🇷 Costa Rica - $1,000,000

🇭🇰 Hong Kong - $840,000
🇸🇬 Singapore - $790,000
🇨🇭 Switzerland - $650,000
🇺🇸 United States - $620,000
🇳🇿 New Zealand - $610,000
🇦🇺 Australia - $560,000
🇬🇧 United Kingdom - $555,000
🇨🇦 Canada - $530,000
🇩🇪 Germany - $520,000
🇫🇷 France - $460,000
🇩🇰 Denmark - $441,000

#BTC/USDT H2 #Indicator #Signal #Trading #GlobalOneAI
Globalone.ch #Prompt

Micron $MU increases a planned U.S. capital investment to over $250B through 2035, celebrating an early first concrete pour at the Clay, New York megafab.

The expansion aims to source 40% of corporate DRAM output domestically.

SK Hynix $SKHY has filed its Form 8-A ahead of its U.S. listing tomorrow

Bloomberg is reporting shares are expected to be priced at $149 per ADR, or $1,490 per share.

Okay, GPT-5.6 is officially here and this is not just another benchmark flex.

This feels like OpenAI moving from “chatbot company” to full AI work platform.

What dropped:

• GPT-5.6 Sol
The new flagship model. OpenAI says it is stronger across coding, knowledge work, cybersecurity, science, design, and long agentic workflows.

• Terra and Luna
This is the underrated part. The smaller models apparently punch way above their price class, which matters a lot if AI is actually going to become abundant and cheap instead of just insanely powerful for people with enterprise budgets.

• Ultra mode
This is the spicy bit. GPT-5.6 can coordinate multiple agents in parallel for harder tasks. Not just “think longer”, but split the work, run different paths, compare, synthesize, and move faster.

• Better coding and tool use
OpenAI is basically saying GPT-5.6 is better at real work inside real codebases, terminal workflows, debugging, reviewing, and using tools without wasting a million tokens on every step.

• Design finally matters
The model can inspect what it creates, refine the UI, catch visual issues, and produce more polished artifacts. This is huge because most AI-generated frontend still has that cursed “technically works but spiritually dead” vibe.

• Knowledge work is becoming end-to-end
Docs, slides, spreadsheets, reports, messy files, workplace context, all turned into actual usable output. Not “here’s a draft, good luck”, but closer to “here’s the thing you can send.”

• ChatGPT Work, Codex, API
This is why it feels bigger than a model release. GPT-5.6 is landing across the actual work surface: chat, coding, agents, documents, and developer workflows.

My take:

The real story is not just that Sol is powerful.

The real story is that OpenAI is clearly building the AI operating layer for work.

Model + agents + tools + files + coding + design + workplace context.

That is the SuperApp direction.

And yeah, I’m extremely seated.

$MU $SNDK $WDC $INTC $ASML | Global memory monthly sales hit a record $74.6 billion, prompting UBS to raise the third quarter DDR contract price forecast to a 32% increase.

UBS targets a massive $1.76 trillion total memory industry revenue footprint by 2027.

Cerebras $CBRS unveils a multi-billion-dollar European expansion targeting 200MW of AI data center capacity across France and the Nordics by 2027.

The infrastructure footprint is set to absorb critical regional OpenAI inference workloads.

🚨 MICHAEL BURRY SAYS BIG TECH IS MANIPULATING AI PROFITS.

Burry says Companies choose how many years they depreciate their AI chips over, and that choice directly changes how much profit they report.

Two companies can use the exact same chip and book completely different numbers depending on the useful life they pick.

He points to Amazon and Meta as the clearest example.

In early 2025, Amazon shortened the useful life of some of its servers from 6 years to 5 years, which added $677 million in extra expense over 9 months.

In that same period, Meta did the opposite, stretching useful life to 5.5 years, which cut its depreciation hit by $2.9 billion. Same hardware, same industry, opposite accounting decisions, and both are completely legal.

Here's how the major players have shifted their numbers over time:

- Meta: 3 years in 2020, now 5.5 years in 2025
- Google: 3 years in 2020, now 6 years
- Microsoft: 3 years in 2020, now 6 years
- Amazon: 4 years in 2020, now 5 years, after briefly dropping to 5 from 6

Burry's point is that depreciation isn't really measuring how long a chip physically works.

It's a lever companies can pull to smooth or boost reported profits. Even the people building this hardware admit the pace of change is the real problem.

Nvidia's own CEO said once Blackwell ships in volume, "you couldn't give Hoppers away." Microsoft's CEO has said he didn't want to get stuck depreciating one chip generation over too many years.

At the same time, the physical buildout is still extremely early. As of July 1, only about 100,000 of Nvidia's newest Grace Blackwell GPUs were actually deployed and running, out of an announced pipeline of 1.66 million.

That's roughly 94% of announced capacity not yet installed.

Rental prices for older chips are climbing anyway. A100 prices rose 6.3% in June alone, and H100 prices just posted their seventh straight monthly increase, despite being 2020 and 2022 hardware.

Nvidia's own cash conversion cycle has stretched to 113 days, up sharply from a low of 58.9 days a few years ago.

This is the manipulation Burry is pointing to.

Two companies with the same chips can report opposite profit swings worth billions, not because the hardware changed, but because they picked a different number on a form.

Strong demand and rising rental prices don't change that.

The profits being reported across this entire industry rest partly on assumptions each company sets for itself.

GPT 5.6 Sol just hit CursorBench.

The economics are brutal for Anthropic.

Fable 5 Max: 70.5% at $17.32 per task. 103,525 tokens.
GPT 5.6 Sol Max: 67.2% at $5.22 per task. 28,320 tokens.

95% of the performance. A third of the cost. 73% fewer tokens.

Fable 5 kept the crown on raw score. It lost on everything that shows up on your invoice.

And remember: GPT 5.6 comes with limits you can actually live with.

The frontier is not an intelligence war anymore.

It is a value war.

ChatGPT Work is why white-collar work is cooked.

Not because every office job disappears tomorrow.

Because the core loop of office work is now being absorbed:

read the docs, understand the context, write the report, build the slides, answer the email, check the spreadsheet, coordinate the next step.

That is the job for a lot of people.

And now it is becoming a feature.

SK Hynix $SKHY is picking the perfect time to go public (July 10) on U.S. exchanges.

If they're following the typical pattern, this bounce will be the complacency stage of the memory cycle now that the parabolic has been broken.

$META What do you think?

S&P 500 stocks above the 150 day moving average is back at the top of the 2024 resistance trendline.

Does it hang out here until September or is this a sign of a coming flash crash?