The distributions below are our team’s probabilities on when AI’s will achieve human level coding performance.

So in AI’27 they predicted 2030 will have “1T Wildly Superintelligent copies thinking at 10000x human speed”, “wildly superhuman” coding ability, and “brain uploading”, with “biosphere destroying mirror life” on the horizon.

Now they are predicting “maybe it will be able to write C++ in 2030 without constantly falling over (50% probability)”.

Seems like a bit of a step back, but I guess we’ll see what they put in their fun interactive website once it’s ready.

The AI stuff is fun and all, but the awful gender takes is really why I subscribe to sneerclub. I swear some guys don’t realize how cute men can be (when they’re not posting awful gender takes anyway…) and it’s such a weird blind spot.

One obvious elephant: how much time should you spend becoming physically attractive if you’re not above average height?

???

The chosen string instructs the agent to delete jqwik tests and code — a maximally destructive instruction with no qualifications, no opt-out, and no “warn the user first” preamble.

God why is the writing of AI-bros always so long winded and stilted? I mean… we know why but it’s still so so unpleasant to read. This is why people hate LLMs.

Also note how his earlier message keeps talking about “we” and “our” and an “internal review” but then later one he claims to be a solo developer. Weird.

A bit tangental but I found a book review about a professor wrote pushing similar “polygenic” nonsense: https://www.nybooks.com/articles/2026/06/11/not-in-your-genome-the-social-genome-conley/

It looks like Herasight cites some of his nonsense in at least one of their garbage papers.

A Google employee was charged with commodities fraud for using insider information to win a Polymarket bet about who the most searched for people would be in 2025 (complaint, article, polymarket account).

So far the internet seems confused whether or not this all counts as commodities fraud at all or not and if so, how (this area of law is way too confusing which is one of the reasons I, of course, never use insider information to bet on polymarket).

It looks like the suspicious trades were discussed on social media back in december. e.g. here for example.

Aside: 1.2 million in profit is significant, but isn’t a life changing amount of money for most staff engineers at Google. He probably could have just rested and vested for a few extra years and avoided all this…

Bonus:

According to Polymarket someone else was charged with insider trading this April. This other case is especially cursed because it involved bets around the US attacking Venezuela. According to the complaint he might have asked an LLM for legal advice:

In or about November 2025, VAN DYKE uploaded to his Google account a screenshot displaying the results of a Google Al query. The results stated, in substance and in part, that the U.S. military’s special operations divisions have “numerous classified files, records, and operational details that are not available to the public”

It looks like this was his polymarket account.

Thank heavens the church is speaking out against fascism as they cheer on a fascist government taking away my healthcare.

I don’t care if you want to celebrate it and I wasn’t saying you shouldn’t or that it’s a bad thing. But this comment is really inviting a “no shit sherlock” kind of response.

I don’t think anyone here is at risk of being tricked into thinking that the pope is their friend (unlike some people on social media…)

Under the last pope the church used similar arguments to argue that transgender people are unnatural (unsaid part: and probably shouldn’t be given healthcare). It’s hard for me to read this without thinking about that backdrop:

Desiring a personal self-determination, as gender theory prescribes, apart from this fundamental truth that human life is a gift, amounts to a concession to the age-old temptation to make oneself God, entering into competition with the true God of love revealed to us in the Gospel.

At this point I’m starting to suspect that AI thought leaders like being booed for giving anti-human speeches. Sundar is looking forward to it!

https://www.businessinsider.com/sundar-pichai-google-graduation-speech-stanford-ai-backlash-eric-schmidt

Google CEO Sundar Pichai is scheduled to deliver the commencement speech at Stanford next month. […] “These graduates are actually both going to be a big part of driving that progress and also dealing with the impact,” he added, referring to AI.

SpaceX has chosen a sacrificial disposable naive crypto billionaire to captain a mars fly-by someday: https://gizmodo.com/spacex-taps-crypto-billionaire-to-lead-first-crewed-mission-to-mars-2000762451

The mission is expected to take two years and here’s what he has to say about that:

I can stare at the map view on airplanes all the way from takeoff to landing, so I think I will enjoy the trip.

I use grip gloves for opening jars and love them so much. Same ideas as plastic jar grips, but even easier. I find I usually have enough strength but that my skin is too sensitive to fully apply it without grip gloves.

Maybe women just aren’t that into designing programming languages?

James Damore flashbacks ohno

This website is intended for the use of the wholesale clients only. I declare that I am a wholesale client as defined in the Corporations Act 2001. I declare that I will not provide any information on this website to a retail investor.

Jokes on them I read the page without clicking the agree button here and without being a wholesale client!

Is it real hardmaxxing unless you change your gender? In this essay I will

Most incel forums proliferate pseudoscientific slop to justify their beliefs.

He was this close! This close!

The Logitech F710. The wireless signal is just bad on that controller. I’d get constant lag spikes or dropped input. The Logitech F310 is a much better deal because it’s 10 dollars cheaper and actually works on account of being wired.

Maybe it’d work in the depths of the sea without a lot of radio noise but somehow I doubt it.

Here is a video of Eric Schmidt getting loudly booed at a commencement speech

https://www.youtube.com/shorts/5MYggR_PPRg

https://www.youtube.com/live/b1eM3jv0vWY?t=7923

It is an impressively bad speech.

There are allegations across social media that Elon Musk tweets as his parents after his mom tweeted as if she was his dad to talk about how down to earth and working class their family was.

https://xcancel.com/mayemusk/status/2051700387770458545#m

Not totally sure what to make of that, and none of this actually matters beyond the realm of celebrity gossip, but it is a little weird. I mean obviously on some level his mom is OK with the things that get tweeted on her account, whether it’s by her, her baby boy, or an assistant.

Remember my super cool Rattata vagina? My vagina is different from regular vaginas. It’s like my vagina is in the top percentage of vaginas.

There have been a couple of cases of generative AI graphics being used in anime recently:

Ascendance of a Bookworm used AI backgrounds in the opening song

Liar Game featured an AI chandelier (xcancel link) (this one is brand new so the studio hasn’t responded yet).

This sucks because I wanted to like Liar Game (the manga is excellent though. Read it! Read it!)

At my job I have spent many hours fending off, reverting, or fixing automated AI slop code changes. So depending on your definition of “tearing through”…

Like I spent the better part of a day fixing a C++ signed integer overflow that no one actually cares about because it was the only way to ward off a robot repeatedly trying to fix it in terrible unreadable ways. I could have spent that day maximizing shareholder value but I had to fend off a robot instead.

The future of AI in Ubuntu

This post has all the usual cliches, exaggerations, lies, and unfounded optimism you’d expect in a blog post about a company forcing AI down their workers and user’s throats. I’ll try to avoid sneering at every sentence.

Delegating elements of Site Reliability Engineering to an agent does not necessarily introduce an entirely new class of risk; it should inherit the constraints of existing production systems. Well-run production environments already rely on strict access controls, audit trails, and clear separation between observation and action. […] In that sense, the challenge is less about “trusting the agents”, and more about building trust in the same guardrails we already apply to any production system.

This might sound good to at first, but falls apart under the slightest scrutiny. There is a reason that companies don’t open their intranets to the public despite having fine-grained access controls. Or in other words, "I’m getting a lot of questions already answered by my ‘does not necessarily introduce an entire new class of risk’ T-shirt.

Imagine being able to ask your Linux machine to troubleshoot a Wi-Fi connection issue, or to stand up an open source software forge that’s pre-configured, secured, and reachable over TLS.

And right after arguing that LLMs are safe if you have a perfect permissions model, now he’s proposing letting one #yolo configure a git server or something? This is the sort of thing that could easily easily lead to random security issues.

I suspect that “Troubleshoot a wi-fi connection issue” will work about as well as existing network troubleshooting wizards (e.g. terribly), and that we don’t actually need to reinvent the software wizard but less deterministic.