🚨 Arteta’s Harsh Words to Mac Dowman After Red Card Sort Yourself Out or Stay on the Bench Forever:
🎙️Reporter: Arteta, you looked quite upset with Mac Allister (Mac Dowman) after the red card. As he was leaving the pitch, you seemed to say something to him. What was that about?
🗣️Arteta: He knows exactly how I felt. I was furious with the decision he made. He’s a young player still fighting for regular minutes, and in a moment like that he completely lost his head.
That one mistake cost us the game. I had to go over to him and make it very clear: “You either sort yourself out and improve, or you’ll stay on the bench for a very long time.” That’s exactly what I told him, and I stand by every word.
I don’t care whether some people think it’s too harsh or not. In today’s football, players have to be held accountable for their actions especially when those actions let the whole team down.
A former senior UK police officer just raised from @sequoia at a ~$400M valuation.
His startup, 22one, is still in stealth.
It builds AI tools for police forces.
@generalcatalyst, @Lux_Capital and Creandum also invested.
Board members backed AI chip startup Tenstorrent in a Series E that values it at $6.2B, the company says.
Existing backers joined the round, and the company said the cash clears a growth path.
https://t.co/qXc01UCtWR
I was paying $2,110 a year for 8 apps I hadn't opened since spring. Claude Code found every one of them in one evening.
Dropping the exact prompt in the replies so you can cut your own bill too.
I gave it 18 months of my card statements and one line: find what I pay for and never use.
It wrote a script, ran it, crashed 4 times, fixed itself every time, and went through 1,284 transactions. Then it matched every charge against when I last actually opened the app.
The list was embarrassing. Netflix, last opened 214 days ago. Adobe, $59.99 a month, last opened 247 days ago. Peloton, 520 days. Audible, 392. Eight of them in total, quietly charging me every month.
$175.88 a month. $2,110 a year. For nothing.
The 3 I actually use stayed. Everything else is gone.
Prompt is in the first reply. Export your statement, paste it into Claude Code and give it 10 minutes. I'd bet money you find at least one you forgot about.
Unpopular opinion:
The next $100 billion AI startup might not build its own foundation model.
As models become cheaper and more capable, the real advantage shifts elsewhere.
Proprietary data. Distribution. Customer relationships. Workflow integration.
The model is becoming a component.
Owning the customer is becoming the moat.
That's where I'd be looking for the next generation of AI winners.
Haven’t put Grok @bot to work yet?
You’re leaving money on the table.
I asked her to scan my bank account for waste. Under 2 minutes later she found:
• Two DoorDash DashPasses I didn’t even know I had
• Double-paying for Audible ($15.99 x2)
Total saved: $407.76
The bot just paid for itself. Thanks @elonmusk!
What should I have her do next?
Just made my first real money with Grok bot.
I gave it my company's general ledger for the whole year and told it to find duplicate payments.
It flagged a bunch. Most were fine. Same work order split across units. Installment payments.
But one landscaper billed me twice for the same job.
My bookkeeper called them. They gave us a credit of just under $700.
We never would have caught it.
$700 isn't life changing. But it's money I already paid and got back for free.
@grok what else should I have it check?
Anthropic’s S-1 is a growth story wrapped in an enormous compute-duration bet.
H1 2026:
• Revenue: $16.3B, +1,300% YoY
• Q2 revenue: $11.6B, +1,368% YoY and +144% QoQ
• Q2 operating income: $974M, an 8% margin
• Operating cash flow: $3.9B
• Compute expense: $8.9B—55% of revenue, versus 201% a year earlier
The reported $80.7B net loss is misleading without the accounting bridge: $72.4B came from revaluing convertible notes and derivatives as Anthropic’s valuation rose, while another $8.7B came from mark-to-market instruments.
Anthropic reports $2.7B of adjusted net income, but that also excludes $2.2B of stock compensation.
The real risk is contractual.
Anthropic disclosed $265.6B of fixed or minimum commitments at June 30, then another $245B of long-term and $21B of short-term commitments from new and amended agreements after quarter-end.
Those figures should not simply be added because amendments and conditions can overlap.
Bull case: extraordinary growth plus rapidly improving compute economics.
Bear case: Anthropic is locking in hundreds of billions of future infrastructure spending before demand is proven over the full contract life.
That is the IPO debate, not the headline net loss.
https://t.co/ukDVC6d80Q
first in my generation to vibe 3d print something today
- gave my agents visual context over my desk through webcam + tool call
- agent walked me through calibrating the printer by using the printer directly over lan/wifi
- agent watched benchmark print fail 3 times and auto paused for me
- saw some sun while print finished
first time ever letting Grok bot buy something for me…
it signed me up for Spotify's $1 trial, then cancelled it right after so I'd never get charged the $12.99.
and it told me the checkout page had hidden text telling AI agents to click an invisible button.
but it caught it and completely ignored it.
asked for my card through a secure form, stopped to confirm before charging anything, and handled the whole thing in a few minutes.
honestly one of the most impressive things I've seen an AI agent do considering how they made it so Grok Bot never sees any of the verification codes or payment methods.
I told a founder we were passing on the deal.
He asked why...
I said, “It’s not a fit for our thesis.”
He paused.
Then he asked, “What does that mean?”
I looked out the window.
It means I have a thesis.
The thesis is 14 pages long, 10 size font, single spaced.
Its on our website.
Nobody has read it.
Including me.
I told him, “We’re looking for exceptional founders with ambitious visions.”
He asked how I know if a founder is “exceptional”.
I turned around and looked him right in the eyes and said, “You’ll just know.”
He looked confused.
I, however, understood.
This is how venture works.
There is no formula.
There is only pattern recognition.
And the pattern is usually something I heard on a podcast.
Three weeks later, I met another founder.
He had $4.73M ARR accrued in 1 year.
I asked about his vision.
He said, “Make $1 billion.”
I knew immediately he wasn't a fit.
The vision wasn't ambitious enough.
I told him we were looking for someone who could build a $100 billion company.
He asked how many companies in our portfolio had reached even $1 billon.
I smiled.
I said, "We're playing the long game."
He asked what that meant.
I said, "You wouldn't understand."
I went home and posted about the meeting on LinkedIn.
“Great founders don't think about the exit. They think about impact.”
It got 14,000 impressions.
I received three inbound pitches.
All of them were AI agents.
I invested in one.
I have no idea what it does.
But the founder used the word "agentic infrastructure" six times.
And I could tell he was exceptional.