While Netflix and Paramount battle over Warner Bros Discovery, the real winners are the bankers. JPMorgan and Allen & Company will collect about $180M in M&A fees no matter who buys the studio. JPMorgan is also raking in another $189M from financing the massive bridge loan used to split Warner Bros into two ahead of the sale. Between advisory and financing fees, JPMorgan alone stands to earn over $280M from the deal, before the ink is even dry. (Reuters)

The Federal Reserve’s independence is on the line as the Supreme Court hears a case over Trump’s attempt to fire Fed Governor Lisa Cook. If the court sides with Trump, presidents could gain sweeping power to remove central bankers who disagree with them, effectively politicizing interest rates. Markets are watching closely: a ruling for Trump would let the White House reshape the Fed, pressure it to cut rates, and potentially rattle global confidence in the U.S. financial system. (CNN)

OpenAI is rolling out an “age prediction” system inside ChatGPT to identify users under 18 and automatically apply stricter content filters. The system looks at signals like stated age, account history, and usage patterns to flag teen accounts. If you’re wrongly labeled a minor, you can verify yourself with a selfie via OpenAI’s partner Persona. The move follows growing scrutiny over AI’s impact on kids and past incidents where ChatGPT served inappropriate content to minors. (TechCrunch)

Meta is asking its independent Oversight Board to help define when public figures (especially politicians) should be permanently banned from Facebook or Instagram. The move follows a case where a high-profile account was indefinitely suspended for violent threats, hate speech, and harassment. Meta says it needs clearer, pre-defined standards to avoid accusations of political bias while still enforcing its rules, especially after the backlash from banning (and later reinstating) Donald Trump. (Social Media Today)

Amazon CEO Andy Jassy confirmed that some Amazon prices are now rising as President Trump’s tariffs work their way through supply chains. Sellers had stockpiled inventory ahead of the tariff rollout, but that buffer is gone, forcing many merchants to either raise prices, eat the cost, or find a middle ground. While Amazon says overall prices remain “within normal fluctuations,” Walmart, Target, and Home Depot are all warning that tariffs will push consumer prices higher in 2026. (CNN)

He open-sourced large parts of 𝕏’s recommendation algorithm (the system that decides what you, me, and roughly 600 million users see every day).

He also admitted something most of us knew already. 

“We are trying to make the algorithm less dumb. It will improve every month.”
— Elon Musk on 𝕏

That line alone tells you everything you need to know about the state of the platform right now.

The algorithm needs a lot of work, but according to Musk, his team is on top of it, and they’re taking user criticism and recommendations seriously. 

If you’ve been on 𝕏 lately, you’ve seen the complaints.

Feeds are getting hijacked by one random click.

Entire timelines are turning into geopolitics, sports, or rage bait. And, great posts from people you follow (and actually care about) are getting buried.

People are even resorting to creating second accounts just to keep their main feeds usable.

This means 𝕏 is still over-weighting short-term engagement (clicks, likes, quick reactions) instead of relationship signals like whether the same people come back, reply again, click your profile, and keep choosing to read you over time.

And that matters for anyone trying to build:

Musk confirmed that xAI (running on 20,000 GPUs at the Colossus data center in Memphis) now powers the recommendation system.

In other words:

Most, if not all feeds are algorithmic now, but what changed on 𝕏 is that a single AI model is now deciding what rises and falls, instead of a bunch of engineers hard-coding what they think people should see. 

The xAI model predicts:

It no longer follows a lengthy list of hand-coded rules; instead, it looks at behavior.

Which is why your feed feels a little strange and chaotic right now. The model is still learning what good engagement looks like.

Posts that trigger real conversation outperform posts that just collect likes.

A thread with 50 thoughtful replies will perform a lot better than one with 5,000 likes and no discussion.

Replies signal: “People stopped scrolling and thought.

𝕏 doesn’t care how many people saw your post. It cares how many stayed.

Posts that get read, re-read, and scrolled slowly get amplified, even if they don’t go viral immediately.

That’s why:

often outperform flashy one-liners.

Anything that sends users off 𝕏 tends to get less reach, similar to platforms like LinkedIn. 

This hasn’t changed.

Which is why you’ve got to:

Never lead with the exit.

Musk has publicly said the algorithm still over-optimizes for “junk engagement.”

That’s a huge admission.

It means 𝕏 will (hopefully) move toward:

This is exactly what happened on LinkedIn and Instagram before their big ranking shifts.

Here’s the translation if you got this far:

If the algorithm overreacts → be careful what you engage with
If replies drive reach → don’t post and disappear
If links are suppressed → move them to comments
If big accounts are favored → comment under them instead of competing
If trust is broken → consistency beats virality

If you want to get on the algorithm’s good side, try this:

The 𝕏 algorithm may still be “dumb.”

But, according to Musk, it will be improving every month now that it’s learning from how humans actually behave.

Worst case?

We all end up back on LinkedIn.

Best case?

𝕏 reverts back to its former ‘Twitter’ glory days everyone misses so much.