Jordan Rogers

Partner @ Co-founder @ RevOps. Done-for-you websites. Utah Jazz. Family. Travel.

[Jordan Rogers @jhmconsult]

I work on two things.

GTM Advisor: RevOps for B2B.
https://t.co/fqBCAHYVOF

RevenueTools: done-for-you websites for contractors.
https://t.co/W1A0ZKZG1r

[Jordan Rogers @jhmconsult]

Went camping this weekend and my sister-in-law blew my mind….

Who freezes their flour for 48 hours after buying to get rid of weevil eggs? 🤯

I’m over here living in the third world with my weevil infested flour apparently.

[Jordan Rogers @jhmconsult]

@StefanoGPT Leaves weekly unchanged

[Jordan Rogers @jhmconsult]

@SpencerJCox - can you please work with @elonmusk and @Tesla to get CyberCabs running in Salt Lake corridor? It is embarrassing the Silicon Slopes nick name watching SF and Texas.

[Jordan Rogers @jhmconsult]

@Jason That AI job loss is real.

CC: @chamath and @DavidSacks

https://t.co/L4K2kOMC18

[Jordan Rogers @jhmconsult]

@Evautoalex @niccruzpatane I want SLC...but you are probably the right man for the job.

[Jordan Rogers @jhmconsult]

@tarasshyn I mean currently 55 on the site and 1.5 million visitors in the week...

[Jordan Rogers @jhmconsult]

Yikes. Take a shot a competitor without naming them. 👀

[Tibo @thsottiaux]

Across the plans Astra will be included in the normal usage allocation and you will be able to use 100% of it towards Astra.

[Jordan Rogers @jhmconsult]

@matteopelleg hmm. test.

[Jordan Rogers @jhmconsult]

Another thing those burning through Fable tokens should note. Don't let the chat go idle (5 minute pause).

Prompt cache is a 40X tax if you idle. A hit refreshes the 5 minute timer. A cache read costs $0.25 / MTok. A fresh input costs $10 / MTok!

Wait 5 min + 1 sec and that same request is 40X the cost.

Keep alive every ~4–5 min, or pay for the 1-hour TTL. New tokens and output still bill full price.

[Jordan Rogers @jhmconsult]

Fable 5.1 strengths: Long-running coding/research agents, knowledge work (GDPval-AA, briefcase-style evals), and the current independent Intelligence Index lead. Cheaper cache reads help multi-hour agent loops. Mythos 5.1 (restricted twin) scores even higher on some terminal coding tests.

Astra strengths (reported): Hard math, science agents, automation, SWE, and especially cybersecurity/capability evals. OpenAI also emphasizes token efficiency vs GPT-5.6 Sol and “price per completed task.”

Caveats: Day-one tables favor the publisher. Cross-lab comparisons are noisy. Fable 5.1 has more public independent coverage so far because it shipped two days earlier. Astra’s cyber numbers come with extra access controls.

[Jordan Rogers @jhmconsult]

I really hope we vote out people like @BernieSanders who are fear mongering AI. Can’t you see this is a space race? If we don’t win in the USA, someone else will.

[Jordan Rogers @jhmconsult]

Yikes. Not Grok too.

[Jordan Rogers @jhmconsult]

So…what do we do when Claude and OpenAI are down at the same time?

Bets on which will give a reset?

[Jordan Rogers @jhmconsult]

@Muddia26Amayri @richardhutton 1. Launch Fable: claude --model fable
2. Turn on high effort: /effort ultracode
3. Force cheap agents: export CLAUDE_CODE_SUBAGENT_MODEL=sonnet

Fable coordinates the big strategy, Sonnet does the heavy lifting. Massive cost savings!

[Jordan Rogers @jhmconsult]

@matt_teeixeira Advice on getting cursor to do better writing? Agree on the coding side. But I often need to do client write ups and Cursor is by far the worst writing.

[Jordan Rogers @jhmconsult]

@OverlyTrev @elonmusk I think @elonmusk has made it clear that @Tesla_Optimus and Cybercap are the priorities ATM.

[Jordan Rogers @jhmconsult]

@JoeRensing @LoganARobison @bot Send me a DM

[Jordan Rogers @jhmconsult]

@poteto - did we just get a @bot reset?!

[Jordan Rogers @jhmconsult]

Fable 5.1 released. Grok 4.7 coming in days. Astra coming "soon." What a time to be building.

[Jordan Rogers @jhmconsult]

@moiiikaaa Codex $200. Claude $200. Grok Heavy $300 (got it when it included Cursor Ultra too).