Jon C

[Jon C @JonCrosby_]

Intelligence is overrated. Hard work, habits and vision compounding over long periods of time proves far more valuable.

[Andrew Yang🧢⬆️🇺🇸 @AndrewYang]

I try not to delight in anyone’s downfall - we’re all human.

[Jon C @JonCrosby_]

Opus 5.5 > Fable 5.2.

Communication 🏆

[Jon C @JonCrosby_]

@DimitrisPapail Sadly, this is accurate. It’s a regression, back to Opus 5.0, let me solve 1 thing but always bring you back 3 more.

[Jon C @JonCrosby_]

@ArtificialAnlys Quick reminder @GeminiApp Carbon never was #1.

Claude's Fable 5.2 beat them to market, both in early and general release.

Please stop rewarding Google's two week early announcements @ArtificialAnlys

[Jon C @JonCrosby_]

@mcuban Mostly agree. We desperately need a more effective mechanism of bringing these ideas to life using fair capitalism (e.g. Cost Plus Drugs, Costco, Vanguard) that drives down cost.

[Jon C @JonCrosby_]

Why is it so hard for politicians to understand? Your voters' top priorities must be YOUR top priorities. The single worst move you can make is retribution, which gets media coverage, but distracts and sends a message you care more about revenge than their priorities. Everything fights for the same limited attention, don't burn a news cycle settling personal scores.

[Jon C @JonCrosby_]

@admcrlsn Correct. Retribution isn’t in my top priorities. I want term limits. I want age limits. I want the deficit fixed. I want health care fixed. Align your dang priorities with the voters.

[Jon C @JonCrosby_]

@ClaudeDevs Claude should not be in the business of storing API keys. There needs to be a degree of separation.

[Jon C @JonCrosby_]

Accountability has to come from we the people. Voters being willing to swing either direction and ignoring the letter next to people's names. Who is exclusively focusing on term limits, age limits, the deficit, corruption, getting money out of politics, and securing social security.

[Jon C @JonCrosby_]

Congrats Democrats. Take this from a lifelong Independent. If you go back to focusing on the wrong things, you will lose the Independent and Latino votes as quickly as you gained them. Focus on term limits, the deficit, getting money out of politics, corruption, and securing social security.

[Jon C @JonCrosby_]

@PeterDiamandis No question, helping people. How can we improve on where we are and how can we bring everyone (who cares to come) along.

[DHH @dhh]

This is also 100% correct. If you're stuck with an org that's refusing to adapt to reality, I have complete sympathy for some shock therapy. Because the alternative is that the business becomes uncompetitive, customers eventually leave, and everyone is fired.

[claire vo 🖤 @clairevo]

prediction: it’s going to be the teams moving slowest, not fastest, on AI that open themselves up to the most cyber risk

risk aversion is a risk itself

[Jon C @JonCrosby_]

It’s BS that companies can have their AA score released prior to general release IMHO. You need standardize this. If Fable 5.5 comes out public PRIOR to Carbon and scores higher, then Carbon never was on top. Fable 5.5 has already been testing with security…. They just aren’t stupid enough to announce the model release until it’s actually released.

[Jon C @JonCrosby_]

@ArtificialAnlys Well done. This matches my real world experience incredibly well. People in the weeds will know this is a significant improvement.

[Padam @Padamn__]

the moat is that i got the dawg in me

[Jon C @JonCrosby_]

If the fastest moving company in the world, with unlimited/unrestricted access to the most advanced frontier models and dozens of cybersecurity experts, can't secure itself against a competent person with an abliterated AI, what chance does the average company have?

This isn't going to be solved easily. The first step is letting go of ego and admitting that competent good guys with AI cannot stop a competent bad guy with AI 100% of the time. And when you are talking about utilities, electrical power, water treatment, military weapon systems, etc...it only takes one. They only have to be right one time out of millions to cause catastrophic damage.

We must fundamentally rethink hardware, protocols, and security at the root level to have any chance of actually solving this. The current attack surface is simply too big for the average company to defend.

[Jon C @JonCrosby_]

Understanding when a problem must be decomposed into smaller pieces, knowing what good looks like, and knowing how to measure "good" are the new currencies of the realm.

[Jon C @JonCrosby_]

Build wildly ambitious things. Intelligence and execution are no longer the bottleneck. We have plenty of builders, but we are starved for people with vision, bold ambition, and the taste to know what is worth building.