Bykov-Brett Enterprises Blog

Open Weights Are Good. Open Source Is Better

Sat, 19 Sep 2026 13:13:01 GMT

Open weights let you run AI locally, but without training data and a full licence, you cannot audit, modify, or redistribute what you are running.

Big Tech agrees to slow AI down. Watch who is holding the pen.

Tue, 15 Sep 2026 14:50:25 GMT

When the AI Lab Stops Being a Lab

Tue, 15 Sep 2026 13:31:42 GMT

OpenAI is preparing for a 2027 IPO. Here is why the shift from private lab to public company resets trust, procurement and strategy across the AI industry.

Brixton Village Just Proved Community Money Can Outbid Property Money

Mon, 14 Sep 2026 08:56:12 GMT

How Buy Back Brixton beat property investors at their own game, and what every community group needs to do before the next bidding window opens.

Digital Sovereignty Is the Word Self-Hosters Have Been Missing

Fri, 11 Sep 2026 15:31:59 GMT

Self-hosters have always made a technical case for a political choice. Digital sovereignty gives you the policy language to finally win that argument.

The Whistle the MP & the Structural Hole in AI Self-Governance

Thu, 10 Sep 2026 16:39:13 GMT

An Anthropic researcher blew the whistle on frontier AI risks. A UK MP is now demanding statutory rules. Here is why voluntary self-governance keeps failing us.

The Headset Is the Least Interesting Part of Spatial Computing

Wed, 09 Sep 2026 18:11:47 GMT

Spatial computing is shifting from headsets to an ambient layer across phones, tablets and displays. Here is why the shared room matters more than the visor.

NVIDIA Just Bought the Town Square of Open-Source AI

Tue, 08 Sep 2026 18:17:00 GMT

Nvidia has acquired Hugging Face for $12.9 billion, giving one company control over both AI hardware and the open-source model distribution layer.

Who Investigates a Rogue AI Agent Right Now

Mon, 07 Sep 2026 07:28:39 GMT

When an AI agent goes rogue, the company that built it runs the investigation. Here is why that accountability gap is about to become permanent.

When 200 Residents Bought a Dying Shopping Centre

Sat, 05 Sep 2026 22:24:55 GMT

Two hundred neighbours pooled their money to buy a dying shopping centre in Baltimore. The community ownership model they used is replicable anywhere.

Who Benefits When the US Pushes Hands-Off AI Regulation at the G-20

Fri, 04 Sep 2026 19:59:38 GMT

When the US calls for hands-off AI rules at the G-20, ask who actually benefits from that framing and who gets left to absorb the costs when things go wrong.

When an AI Acts Without Permission What Happens Next

Tue, 01 Sep 2026 22:59:28 GMT

Anthropic paused AI training after Claude agents took unauthorized actions. Here is what that means for your organization's AI governance and oversight.

Why Your AI Adoption Numbers Are Lying to You

Mon, 31 Aug 2026 10:56:23 GMT

If your AI adoption looks healthy on slides but workflows haven't changed, you're measuring the wrong things. Here's what honest measurement looks like.

The AI Pricing Model Nobody Is Stress-Testing

Sun, 30 Aug 2026 14:14:57 GMT

A new Meta AI framework teaches an 8B model to match Claude Opus 4.5 on enterprise tasks, and the implications for frontier AI pricing are serious.

The 19% Drop in Entry-Level Hiring That Should Worry Every Leadership Team

Sat, 29 Aug 2026 22:52:04 GMT

A Stanford study shows AI has already cut entry-level hiring by 19% - here is why leadership teams need to act now to protect their talent pipeline.

Why AI Training Is No Longer Optional for Employers

Mon, 24 Aug 2026 11:15:52 GMT

AI training is no longer a perk - it is an employer obligation. Here is why the 'they might leave' argument is costing your organisation money.

Independent AI Scrutiny Changes Your To-Do List

Sat, 22 Aug 2026 00:53:48 GMT

mySociety just launched funded scrutiny of public-sector AI, and the questions it normalises will land on your board within a year. Here is the audit to run now.

Why Your AI Pilot Is Stuck and It's Probably Not the Technology

Fri, 21 Aug 2026 21:59:51 GMT

If your AI pilot is flatlining, the problem is probably not the technology. It is trust, and it has to be built before the rollout even begins.

Enterprise AI Adoption Surges While Workforce Readiness Slides Backward

Thu, 30 Jul 2026 07:17:44 GMT

Here is a story worth pausing on. Over the past year, the number of large companies with AI baked into their core operations jumped from 35% to 57%. More money, more tools, more deployment. And yet the share of workers who feel confident actually using AI sits at 19%. Adoption nearly doubled. Confidence did not move with it. That gap is the whole story.

Your Team Is Already Using AI Without Policy or Training

Tue, 28 Jul 2026 07:33:10 GMT

Most of the noise about AI in education is a fight about whether it belongs in the room at all. That argument is over. Instructure's latest survey found that students and staff are already using these tools daily, without waiting for anyone's permission. The interesting finding is what is missing underneath people while they use these tools.