Fred Marks

Building with cutting edge ai since dec 2022

[Fred Marks @MarksFred]

@GeminiApp nice remote browser feature! Paired with Carbon, chef's kiss!

@OfficialLoganK
@DeryaTR_

[Fred Marks @MarksFred]

@owendesign Jay you are a good man!

[Fred Marks @MarksFred]

@owendesign @Voxyz_ai @GeminiApp Carbon is great orchestrator @willcb

[Rachel Tobac @RachelTobac]

Criminals may attempt to leverage the Outlook outage to phish you. You may receive emails once it’s back up asking you to change your password to prevent further disruption, texts/calls that claim they are IT & need your creds/codes to get you back online. Stay politely paranoid.

[Fred Marks @MarksFred]

@Andrew_Blumson Nice yours arrived?

[Fred Marks @MarksFred]

@Natsecjeff And the ignorant think it’s not an attack.

[Fred Marks @MarksFred]

@Angaisb_ I had 4 subscriptions

[Fred Marks @MarksFred]

@IntCyberDigest @BrianRoemmele

[Fred Marks @MarksFred]

@GeminiApp Carbon gives funny messages like this

[Fred Marks @MarksFred]

@Scobleizer He was cocky like Tiger Woods. Watch his interviews vs pre 2023.

[Fred Marks @MarksFred]

2027 is the year of AI Engines.

[will brown @willcb]

2027 is gonna be about engines

[Fred Marks @MarksFred]

@Scobleizer @ryanlpeterman Robert, brilliant!

[Fred Marks @MarksFred]

@BrianRoemmele Brian, thank you!

[Nous Research @NousResearch]

Introducing Arete-27B: Qwen3.8-27B, post-trained to reinforce Western values.

Chinese open models are currently the strongest available, but they ship with state-aligned behavior in the weights. We built an 803-prompt eval across seven value categories, including freedom of expression, democracy, and privacy, and measured how often a model adopts China’s party line.

After fine-tuning on roughly 53 million tokens, Arete-27B's propaganda rate fell from 22% to 1%, in line with the Western models we tested. Its Hermes Bench score shows no significant performance tradeoffs.

The values in an open model are one fine-tune away.

Our eval, weights, and training data are available now along with the full blog post: https://t.co/iAOANQTVZ6

[Brian Roemmele @BrianRoemmele]

This Grok Bot tests if an online store is optimized for Grok Bot.

It tests 20 points that allow for smooth commerce and sets a rating of A-D on being Grok Bot ready.

Although Grok Bot can overcome most problems, a “D” rating represents nearly a full block to conclude a sale.

I built this Grok Bot for online store owners to be sure the can offer the best Bot based experience. It will show how to optimize a site and drive the changes.

It is also useful for anyone that plans to use Grok Bot at a favorite online store. It will test sites directly and offer ways to use the site or unfortunately for the site, offer alternatives.

Get it today: https://t.co/sz6C7PrwD2

[Kurt Elster @kurtinc]

I turned this 90s clear phone into my house's voice assistant with @home_assistant. You pick it up, and an “operator” answers that accepts voice commands like Alexa. "set the thermostat, change the lights, arm alarm, etc... Plus I added easter eggs. Like dialing 9 plays Liam Neeson's Taken monologue 🤓

[Fred Marks @MarksFred]

@Natsecjeff Keep the faith.

[Fred Marks @MarksFred]

@AestherML Qwen 3.8 27B can’t breathe underwater.

I’m a fan of Qwen.

[Fred Marks @MarksFred]

@Natsecjeff How long will that stock last you?

[Fred Marks @MarksFred]

@ItaMilRadar Is this a recent activity or commonplace?