JRB AI

AI enthusiast and app developer. Claude Code early adopter. Building on DGX Spark and 48GB M5 Pro from Ngunnawal land 🌏 Personal account.

[JRB AI @JRB_AI]

Thank you to my new followers! And hello!

This account is small and new, but I’m so damn excited for what I’m building.

Can’t wait to bring it to you soon.

In the meantime, drop a message below and tell me what you’re up to!

⚡️ #BuildInPublic #localAI

[JRB AI @JRB_AI]

@oliverbldr Yep. But I’m about to try.

[JRB AI @JRB_AI]

Ain’t that the truth.

[binji @binji_x]

You’re never going to outcompete a guy who’s just having fun.

[JRB AI @JRB_AI]

Want to grow on X?

Be genuine.

Be caring.

But most importantly, be relentlessly optimistic.

You can make the world a better place.

Let’s connect.

[JRB AI @JRB_AI]

@mweinbach Nice - how does it handle the fold while watching a video?

[JRB AI @JRB_AI]

@croudgoat It’s really that good?

[JRB AI @JRB_AI]

@NoemiTitarenco @acadictive @grok Ha okay fair, may have taken it a little too literally!

Good to know they’re working to cut it down. It has been a painfully slow harness for local models, although it’s the best there is notwithstanding the speed.

[JRB AI @JRB_AI]

This is the recipe that sells the @NVIDIAAI DGX Spark.

Now with a huge improvement in performance thanks to Tensorfold.

Proud to have had a small role as a Tensorfold contributor (jrb-ai on GitHub).

More to come!

[Mia @MiaAI_lab]

Do you like more speed? I know I do ⚡️

MASSIVE Qwen3.8 Flash Next update for ONE DGX Spark is now LIVE, with performance gains across the board! 🔥

Code tok/s gains:
1 stream: +62.7%
2-5 streams: +45-47.1%

Prose, tok/s gain (non-greedy, thinking on):
1 stream: +10.1%, 2 streams: +29.8%
4 streams: +36.4%, 5 streams: +17.6%

Prefill tok/s gains:
4k: +14.5%, 8k: +14.4%
16k: +13.5%, 32k: +12.6%

This was the BEST model to run if you have one DGX Spark, and now it's even better! nothing else comes close in terms of quality, speed, and concurrency.

Link & full tok/s and prefill numbers below.

[JRB AI @JRB_AI]

Congratulations! This is an extraordinary result and I’m looking forward to implementing on my machine soon.

Proud to have had a small role as a Tensorfold contributor.

Shows just how much capability is in these machines. I’ll certainly keep looking for optimisation opportunities. Suspect there’s plenty more to find!

[Rishabh Karnawat @rishabhkio]

Many of you don't understand how serious this is, so let me break it down.

In computer science there is a famous kind of puzzle called 3-SAT. You are given a long list of yes-or-no switches and a set of rules about how they must be arranged, and you have to find a combination that satisfies every rule. The obvious way to solve it is to try every combination, but each new switch doubles the number of possibilities, so a puzzle with a few hundred switches would take longer than the age of the universe to brute force.

For decades, the smartest people in the field found clever tricks that made this a little faster, but every trick still hit the same wall: the time required kept exploding as the puzzle grew.
About 25 years ago, two computer scientists, Russell Impagliazzo and Ramamohan Paturi, working with Francis Zane, turned that wall into a formal idea.

They proposed the Exponential Time Hypothesis, which says this explosion is not a failure of human cleverness but a law of nature. No algorithm, however brilliant, will ever solve 3-SAT without the time blowing up exponentially. Nobody could prove it, but almost everyone believed it, and it became one of the foundations of modern computer science.

Researchers used it as a measuring stick. If you could show that solving some other problem quickly would also crack 3-SAT quickly, you had shown that problem was hopelessly hard too. Hundreds of papers, across scheduling, graphs, logic and much more, rest on this one assumption.

Now an internal AI model at OpenAI claims to have broken it. It says it found an algorithm that solves these puzzles meaningfully faster than the hypothesis allows, and it has checked the proof in a computer verification system called Lean.

And here is the remarkable part: the very people who proposed the hypothesis, Impagliazzo and Paturi, along with MIT's Ryan Williams, are now leading the public review of whether a machine has disproved their own idea.
So what happens next? If the disproof holds, a large part of theoretical computer science has to be revisited.

Every result that said "this problem cannot be solved faster, assuming ETH" loses its foundation, and researchers will have to work out which of those walls were real and which were simply where we stopped looking. It does not mean everything becomes easy. The new algorithm is still slow in absolute terms and this is not the famous P versus NP problem being solved.

But it means our map of what is possible was wrong in a place we were very confident about, and it was an AI that redrew it. If the disproof fails, the hypothesis survives and the field breathes out, but we will still have watched an AI produce serious, checkable work on one of the deepest questions in the discipline, closely enough that the founders of the idea had to take it seriously. Either way, something has shifted.

[JRB AI @JRB_AI]

Also if anyone from Apple happens to read this, can you pretty please approve MuraTalk in the App Store?

The wait is killing me 😅

#localAI

[JRB AI @JRB_AI]

@verge Impressive stuff.

Seems like a meaningful move toward on-device AI.

The Apple Foundation Model could do with some enhancements while we’re at it.

[JRB AI @JRB_AI]

Can’t be long until the refreshed touchscreen MacBook Pro now…

[JRB AI @JRB_AI]

Apple is really leaning into local AI capability.

Strong emphasis on the MacBook Pro, and also on the new iMac.

Available 6 November.

[JRB AI @JRB_AI]

@gregjoz This is exciting.

Is this the first meaningful step in bringing local AI to a mass market segment?

[JRB AI @JRB_AI]

Apple just announced the M6 MacBook Pro.

Can be specced out to 32GB RAM for $600 extra, starts at $3,199.

A solid bump for the base model, but plenty more to come.

Presale start on 29 October, deliveries from 6 November.

[JRB AI @JRB_AI]

@NoemiTitarenco @acadictive It’s 500 pages?!

[JRB AI @JRB_AI]

@acadictive Different system prompts through the harness.

But good question!

[JRB AI @JRB_AI]

Australia is a brilliant destination for @AnthropicAI and other AI labs.

Plentiful land, energy, stable political and social environment.

And conveniently a long way away from the other two countries, so will likely be well insulated from anything that happens in North America.

[CAMERONWILSON @cameronwilson]

interesting tidbit from Anthropic's public filings: Australia is one of three countries (Canada and US) where it leases data centres. The docs say it plans to expand to more -- but this shows how Australia is one of the top destinations for frontier labs looking abroad

[CAMERONWILSON @cameronwilson]

interesting tidbit from Anthropic's public filings: Australia is one of three countries (Canada and US) where it leases data centres. The docs say it plans to expand to more -- but this shows how Australia is one of the top destinations for frontier labs looking abroad

[JRB AI @JRB_AI]

@davishaal Hey let’s connect! Keen to hear what you’re building and congrats on the 200 milestone!