Mikel

always finding another problem to solve

[Open Source Intel @Osint613]

REPORTER: “Last month you said Iranian leaders were very rational people, nice to deal with, strong and smart. Today you said they're scum, sick people, and being led by sick people. What changed?”

TRUMP: “I got to know them”

[Mikel @MikelLabs]

@borrasque_ compute

[Param @ParamSiddh]

1) I switched to AI Engineering 1 years ago!
It was the best career move I ever made.

If you want to start today, here's a roadmap:

[Mikel @MikelLabs]

@TheDreadCEO open your dms

[John Stossel @JohnStossel]

.@AOC and @BernieSanders say AI will take away jobs!

But new technology leads to NEW jobs.

Economist Don Boudreaux: "If we had today the same technology we had 40 years ago, the only web designers would have eight legs."

For more:

[Chubby♨️ @kimmonismus]

Holy moly, Anthropic is getting very serious about recursive self-improvement!

One word: acceleration.

Insane blog article.

Tl;dr:

•We are close to an AI capable of fully autonomously designing and building its own successor

•They stress this isn’t here yet and isn’t inevitable, but could arrive sooner than most institutions are ready for

•Anthropic engineers now ship on average 8x as much code per quarter as they did in 2021–2025

•Task length AI can reliably complete is doubling roughly every 4 months (up from every 7 months)

•Opus 3 (Mar 2024) handled ~4-minute tasks; Sonnet 3.7 (a year later) ~90-minute tasks; Opus 4.6 (a year after that) 12-hour tasks

•SWE-bench went from low single digits to saturated in two years; CORE-bench (research reproduction) went ~20% to saturated in 15 months

•METR found Claude Mythos Preview could work “at least” 16 hours, at the top of what they can currently measure

•As of May 2026, Claude authored 80%+ of code merged into Anthropic’s codebase (low single digits before Claude Code launched in Feb 2025)

•A March 2026 poll of 130 research staff: median respondent estimated ~4x output with Mythos Preview

•One April 2026 example: Claude shipped 800+ fixes cutting a class of API errors 1,000x, work an engineer estimated would have taken a human four years

•Claude-written code quality: worse than human in late 2025, roughly at parity now, expected to be strictly better within the year

•On the hardest open-ended tasks, Claude’s success rate hit 76% in May 2026, up 50 points in six months

•Code-speedup test: Opus 4 averaged ~3x speedup (May 2025), Mythos Preview ~52x (April 2026); a skilled human needs 4–8 hours to hit 4x

•In an AI-safety research project, Claude agents recovered 97% of a performance gap (vs ~23% for two human researchers in a week), over 800 compute-hours and ~$18K

•On picking the better “next step” in research sessions, the best model beat the human choice 51% (Nov 2025, Opus 4.5) rising to 64% (April 2026, Mythos Preview)

•Human comparative advantage, for now: research taste and judgment, i.e. choosing which problems matter and when an approach is a dead end

Three possible futures

•The trend stalls (S-curve), but today’s capabilities still diffuse widely; they consider this least likely

•Compounding efficiency gains, with humans still setting direction; 100-person firms doing the work of 10,000+; they think this is the likely path

•Full recursive self-improvement, where AI builds its successors and pace is set by compute; the alignment outcome here is what they’re least certain about

[levent @__alpoge__]

just, please read this and think hard

[Nate Cavanaugh @natecavanaugh]

Today we're announcing @Special and a new financing led by @a16z.

Our DOGE teammates also invested: Human Capital, Steve Davis, Antonio Gracias, Anthony Armstrong, plus @brian_armstrong from Coinbase and @ssankar from Palantir.

DOGE for the private sector is here.

[John Stossel @JohnStossel]

Young socialists say: we don’t need capitalism—innovation thrives without profit.

One points out that the Soviet Union made a mobile phone.

But how many of you are reading this on a socialist-made phone today?

Idiocy from bizarrely popular social media starts here:

[Reads with Ravi @readswithravi]

The people at the top are collaborating.

[Mustafa @oprydai]

be a full stack thinker.

not just full stack in software.
full stack in reality.

• understand physics → what the world allows
• understand code → what machines can execute
• understand economics → what can survive
• understand psychology → what people actually do
• understand systems → how everything compounds or breaks

specialists optimize one layer.

full stack thinkers see the whole machine.

[Adrian Dittmann @AdrianDittmann]

Neurodivergents are like creative computers, give them enough information and a proper goal, and the results will be remarkable. Palantir understands this.

[Mo @atmoio]

AI is not for you

[Sean Frank @Seanfrank]

PROPAGANDA IM PUSHING:
- it’s the best time to start a business ever
- have a kid. It’s hard but way cooler than having a dog
- take the wonder drugs. Glp1s are cool.
- walk 13,000 steps a day. It’s easy.
- you can reach any successful person on earth just by posting and trying to find their email for 10 minutes
- meta ads are working the best they have in 4 years
- the world needs more podcasts, not less
- the world needs more creators, not less
- we have already reached the age of abundance
- YouTube is finally a good ad channel
- things are about to get even better
- Tik tok shop isn’t a big sales channel, but the ability to engage thousands of motivated creators will 2x your business by sheer amount of ads
- sauna and cold plunge combo is great for mental health
- think more, it’s better than reading
- you can’t be a millionaire overnight, but you can be a billionaire in a decade.
- get married.
- attention is everything. Good or bad, attention is everything.
- play a physical sport once a week. It’s the only way to get friends together as we get old
- do stuff now that’s high risk, it’s okay to look stupid
- there is a billion dollars in your computer, and 100m in your phone
- want to do cool shit? Ask. Anyone will work with you as long as you have a good plan and can deliver

GO DO SOME GOOD

[alexei @alexeixbt]

be retarded
embarrass yourself
stop giving a f*ck
speak your mind
be chalant
be free

[Big Brain Business @BigBrainBizness]

"Facebook wasn't the first thing I built."

Mark Zuckerberg on the failed projects nobody remembers:

[Siddharth @siddharthwv]

This is success:

Make wifi money. Go out at 3pm on a Monday. Take risks. Study the greats. Sell shamelessly. Play a sport. Listen to old songs. Don't believe your negative thoughts. Reading more. Walking more. Scrolling less.

[spor @sporadica]

type of guy to like a post then click “not interested in this post” just to fuck with the algorithm like haha buddy you’re not figuring me out just yet

[Sam Lambert @samlambert]

400 mg of caffeine, 20 mg of adderall, and 3 agents on the go.

[𝐂𝐡𝐥𝐨𝐫𝐢𝐬 @floralholicx]

my goal