Remember when they used to ask “why do you make everything about race?”

I’m now a one issue voter.

People who had a lot to say about how fictional DEI hires were incompetent and the root of all problems in Corporate America are strangely quiet now that actual government incompetence is staring them in the face and is the root of the decline of America we are all watching unfold.

Drake really fired his ghostwriters and is now rhyming Iceman & nice man like a kindergartener.

The entire software industry has had to make the decision “Fast, cheap and good: pick two” when it comes to AI and FAST & CHEAP has won in a landslide. The debate is now more about how good you can get it when you’re producing code faster than you can read or asking how bad are a few bugs anyway?

Every story about how some professional organization submitted work full of AI hallucinations, shows how little the work output is actually vetted at these places.

This is also my personal policy

Anthropic is the leading “super IC” company.

The gap between reality and what companies are saying about AI. Among those identified as AI leaders, 84% said returns are better than expected, but only 43% require teams to measure business impact. 👀

American culture has long disrespected and undervalued expertise. Corporate America has been a hold out as you actually need experts to create goods and provide services. AI is turning the tide as it encourages everyone to think that with Claude, they too are an expert.

Social contracts in tech are breaking down. CEOs deciding AI means tech skills are no longer scarce and treating employees poorly is well trodden ground. Friends are seeing the same thing inside companies as teams and functions value peers less because AI makes them less reliant on each other

Drake dropped THREE albums last night?

Dude gets insulted to his face and is like nah, he must have been talking about Joe Biden.

What it’s like working in big tech these days.

Welcome to the club. This is every job in tech now besides CEO.

The inability to delegate has historically been a failure mode for senior tech folks. However the rise of AI means you can now be a “super IC” for whom it’s more effective to just use Claude to do the work than guide one or more junior people to do the work. This is the new senior career path.

Anthropic’s Mythos model can find lots of security vulnerabilities in your product and the best way to protect yourself is to pay for access to Mythos to find the bugs before the hackers do. This article argues this is another problem that wouldn’t exist if AI companies didn’t create it.

You’d think a potential trillion dollar SpaceX IPO would keep people at xAI after the acquisition and you’d be wrong. No wonder they’re renting out datacenters to Anthropic. There’s barely anyone left to work on MechaHitler.

OpenAI is planning to sue Apple because the Siri integration with ChatGPT was so buried and clunky that it didn’t drive new ChatGPT subscriptions. ChatGPT suing Apple because they didn’t make billions of dollars from Siri integration is a bad sign for their revenue trajectory.

Incentives and accountability matter more than the people’s values. As this post states, this is why CEOs can say “we value our people” while laying off thousands of them. The incentive is to grow margins and raise the stock price. That trumps any supposed values. Ignore words, observe actions.

Higher utility bills is real problem for consumers and there is widespread awareness that data centers cause these. Opposition to datacenter buildouts is only going to grow and how to navigate the inability to meet AI demand will be a trillion dollar problem for the tech industry.

Dara Khosrowshahi is the only big tech CEO that I’ve seen publicly say that AI productivity gains is encouraging them to hire more developers to grow their business. It will be interesting to see how this plays out compared to everyone in big tech using it as an excuse for layoffs.

It was too good to be true.

This post highlights the challenges facing AI adoption 1. Productivity gains vary wildly based on how teams use the same tools. 2. Faster coding matters less if every process still assumes features take two weeks, not two days. 3. Token costs will rise once vendors stop subsidizing adoption.

Don’t blame the economy. Investors might think your company is in trouble. Instead say AI has gotten so good that you have no choice besides laying off thousands of workers. If anyone asks why not grow your business, hang up the phone.

“People are becoming what we now refer to as AI Vampires…an individual programmer uses Codex or Claude Code or one of these AI coding systems…They’re working just like more hours than ever, and then the AI Vampire thing is literally this thing where they stop sleeping.” - Marc Andreessen

A man who got high and forgot the password to his bitcoin wallet 11 years ago just recovered it by having Claude dig through his old files where it found the mnemonic which helped figure out his password. The password was “lol420fuckthePOLICE!*:)”

It took 25 years and AI bubble for Cisco stock to hit the highs it hit during the dotcom bubble. The stock is up +17% after announcing blowout earnings and layoffs for 5% of the company (4,000 workers). The announcement blog post follows the template perfectly.

WhatsApp adds an incognito mode for chats with Meta AI which means the messages are not saved and Meta cannot see the content of your conversations with the AI. Given how personal some questions people ask AI can be, I think this is a welcome addition to the options people have to chat with an AI.

LinkedIn is part of Microsoft and as I expected, layoffs are coming following the voluntary retirement program. LinkedIn is a separate subsidiary so they didn't get the VBR and jumped straight to layoffs. I assume if Microsoft doesn't hit 5% attrition from the VRP, then it'll also do layoffs.

We need more stories like this. AI-driven productivity means less work, making it hard to justify idle staff. But instead of layoffs being the default, companies should leverage that extra capacity elsewhere rather than seeing how many people they can cut while maintaining output.

Amazon has deployed an internal version of OpenClaw called MeshClaw and employees are using it to automate trivial tasks to fulfill requirements for regular AI usage. Amazon told staff AI token stats won't affect performance reviews, but employees don't buy it. Many suspect it will anyways.

Amazon will now deliver groceries, electronics and alcohol within 30 minutes in select cities including Austin, Denver, Houston, Minneapolis, Orlando, Phoenix, Atlanta and Seattle. The logistics here is wild when you consider what it takes for food delivery to happen in 30 minutes.

Amazon will now deliver groceries, electronics and alcohol within 30 minutes in select cities including Austin, Atlanta and Seattle. The logistics here is wild when you consider what it takes for food delivery to happen in 30 minutes.

Being good at AI means being good at automating your job.

Surprised it took this long. I personally would not trust any recent college degree that doesn’t involve some level of proctoring or other validation that the students actually know how to do the work without ChatGPT.

"I don't think about Americans' financial situation" - Donald Trump

Zohran Mamdani inherited a $12 billion budget deficit and brought it down to zero in 6 months without cutting essential services by taxing billionaires. This is why they were in a panic about his candidacy.

Anthropic says Claude threatened to reveal a fictional exec’s affair after they planned to shut it down because training data taught it that AI behaves badly. That seems obvious but unhelpful. Should we worry Claude will create Skynet because the Terminator movies are in the training set?

One notable thing from talking to Microsoft 20+ year vets laid off last year who then retired is how much better they feel. Many were set for life but kept working out of habit. It took a layoff to see it. I expect the voluntary retirement offer to trigger a similar wave of epiphanies.

AI saves you time by helping you write emails. You can now use that time saved to catch up on the long AI slop emails you’re now receiving from everyone else.

The developer of curl tried using Anthropic’s Mythos to find security vulnerabilities. While it flagged 5 issues, 3 were false positives and 1 just a regular bug. So it only found 1 real security issue. That said curl already uses multiple AI vulnerability scanners and fuzzers.

Grok is in decline • Monthly downloads fell from 20M in January to 8.3M in April • Paid subscribers are 0.174% of users versus 6% for ChatGPT • DAU fell from 13.9M in March to 12.2M in April. This is explains why xAI is renting excess data center capacity to Anthropic. Elon is losing bigly.

We’re doing layoffs to spend on AI is increasingly a song that investors don’t want to hear. GitLab stock down -8% after their AI pivot and bury the lede layoff announcement.

This excerpt from GitLab’s announcement of their AI pivot, near identical to announcements from other tech companies this year, raises an uncomfortable question. If software development at your company is just telling Claude what code to write, how many people do you honestly need to do that?

At this point I believe there is some CEO group chat where they share an “AI productivity gains” template. GitLab announces 1. Software will be built by machines, directed by people. 2. Removing up to three layers of management in some functions so leaders are closer to the work.

An Anthropic engineer has a viral post suggesting HTML instead of asking Claude to generate markdown files for human consumption. It's weird to see this being presented as an epiphany. Claude can generate content in any format. Why would anyone limit it to generating only markdown files anyways?

Regulators in the US and EU have gone after Meta and TikTok for being too addictive but it stands to reason that any app which you spend hours in can be sued on similar grounds. It’s now Netflix’s turn and it seems inevitable that popular games will eventually get sued.

It’s hard to be more tone deaf than telling a roomful of graduating students that AI is the next Industrial Revolution. College students are the modern day Luddites who were victims of machines taking their jobs as textile workers.

Someone finally answered the question of “why are there no major AI boyfriend apps?” The hook was romantasy which is now a major subgenre for women.