HNDRX Digital

AI architecture, agents & data transformation for businesses that want results (not hype). Practical frameworks | Real implementations.
DM for strategy chats.

[HNDRX Digital @HNDRX_Digital]

CRAZY: we got a quote for a promo video package from a familiar brand.

£10K for the video +£15K and they hand you 50+ influencers who will push it, repost it, and flood the comments so the views go above 500k.

That’s the exact recipe behind almost every “viral” video.

[HNDRX Digital @HNDRX_Digital]

@BillAckman Heartbreaking @BillAckman
The blessing here is that Lucy is still alive.
Others will also benefit from your generosity.
Kudos to you.

[HNDRX Digital @HNDRX_Digital]

@_sun_yi $200k is a Director-level sign-off.
The CEO doesn’t even see it.

[HNDRX Digital @HNDRX_Digital]

@_sun_yi 7.
Not an expert though.

[HNDRX Digital @HNDRX_Digital]

BREAKING: Claude models will now have invisible watermarks embedded in all text, and all metadata attached to files.

#AI

[Chaz Wargnier ☄️ @DataChaz]

How AI market share has changed since ChatGPT launched in Nov '22 😲

[Alexandr Wang @alexandr_wang]

1/ big announcement today: we will be releasing an open weight version of muse spark 1.2 soon.

we also are releasing muse glimmer, a 30B agentic model with open weights under apache 2.0. muse glimmer can run on 24GB of VRAM without losing agentic reliability. 🧵

[HNDRX Digital @HNDRX_Digital]

This is how @edzitron explains AI finances in one video.

[HNDRX Digital @HNDRX_Digital]

Not a talent loss, just smart capital allocation, they profit from the infra either way.

[HNDRX Digital @HNDRX_Digital]

Google’s chief scientist Jeff Dean + 3 top researchers just left after decades to found Discovery Loop.
Alphabet’s response? Become a founding investor and supply their compute for the first year.

[Anthropic @AnthropicAI]

We asked an unreleased research version of Claude to take a stab at the Riemann hypothesis.

It didn’t solve it, but it did make strides on a related problem: it increased the lower bound for the fraction of zeros of the Riemann zeta function that satisfy the hypothesis from 41.6% to 67.2%.

https://t.co/aZDvqqhHRi

[HNDRX Digital @HNDRX_Digital]

BREACH DETECTED: In Australia's first known autonomous AI cyberattack, an OpenClaw agent used a vulnerability in a gym’s API to leapfrog scheduling restrictions for a gym class, and then forcefully cancelled another person's reservation to move its user up the list, per ABC.

[HNDRX Digital @HNDRX_Digital]

@plumberbutt97 Better for?

[HNDRX Digital @HNDRX_Digital]

The future of AI security may depend on the thing regulators want to restrict:

Open models.

Hugging Face used an open model to defend against a breach, running privately on their own infra.

Open weights aren’t just a risk. They’re a defence capability.

@ClementDelangue

[HNDRX Digital @HNDRX_Digital]

10/
The biggest AI opportunity isn’t using better models. It’s making more real-world work verifiable. Whoever builds those evaluation systems will shape the next decade of AI.

[HNDRX Digital @HNDRX_Digital]

Karpathy described the highest-paid skill in #AI engineering, and it is not prompt engineering or mastering the latest model. It’s something far more durable: verifiability.

[HNDRX Digital @HNDRX_Digital]

9/
That’s a rare skill because it requires technical ability, business understanding and clear thinking. You have to define what “good” actually looks like.