Marc Ryan

Insights / Product / Data Guy. Playing with Ad-Tech, Decentralization, Privacy, Decision Intelligence, M/L and agentic AI. Posts are my opinions.

[Marc Ryan @marcryan]

🤯@cursor_ai just Rick Rolled me. I gave it permission to use the browser to check YouTube for a project and of course the video it pulled up had to be...

https://t.co/kVpb8pcXdi

[Marc Ryan @marcryan]

@mainak @Verizon It's down out in the suburbs as well

[Marc Ryan @marcryan]

@robleathern Mullvad is time and again proven to be one of the better "safer" options. Personally I use @ProtonVPN as it comes bundled in their subscription. Maybe not as bullet proof as Mullvad but much better than other options.

[Marc Ryan @marcryan]

@rohanpaul_ai Yeah. But doing so tempts Roko’s basilisk

[Marc Ryan @marcryan]

@NVIDIAGeForce GeForce Day!

[Marc Ryan @marcryan]

@chipro Maybe try one of these tactics: https://t.co/8KuhJ1YuAt

[Marc Ryan @marcryan]

@chipro Not possible. Just another sign AI is attempting to subvert the human condition😀, we'll all end up dying from a brain bleed after the 600th iteration of failed system prompting.

[Marc Ryan @marcryan]

@emollick Maybe this is a future career path for us mortals bound to be replaced by AI . Writing tests to evaluate AI systems sounds like an in demand skill. 😁

[Marc Ryan @marcryan]

We don't do a great job learning from the past and will likely repeat it. The big Q: Will we create an open, interoperable AI operating system or just more walled gardens? What do you think? Drop your thoughts below!

[Marc Ryan @marcryan]

Every day we see new advancements in LLMs that reinforce what @karpathy suggests -- that LLMs are the new operating system. But is this progress or just new semi-monopolies in disguise? Thread below...

[Marc Ryan @marcryan]

So, if LLMs are the new OS, what transparency requirements, system, or governance will stop giants from killing innovation & favoring their tools? How do we build a diverse market? Most AI companies aim to "own" users.

[Ian Nuttall @iannuttall]

Claude Code pro tip:

Ask it to use Gemini CLI with its 1M context window and free plan in non-interactive mode to research your codebase, find bugs, and build a plan for Claude to action

Prompt and link below:

[Dushyant @DevDminGod]

Evil genius plan.. if you have Gemini access, use that to do summaries from within Claude Code (use Geminis massive context window)

[Marc Ryan @marcryan]

I used to think governance was just corporate red tape until I saw a messy dataset tank a product launch. Shelly's idea of an AI Governance Manifesto is solid. If your data’s a dumpster fire, no amount of AI magic will save you. Things to do: set clear ownership rules, automate compliance checks, and train your team to spot bad data fast. Learned this integrating acquisitions; one wrong move, and you’re toast.

https://t.co/ThKTnvjC0j

[Shelly Palmer @shellypalmer]

In the race to harness AI’s transformative potential, success stories are emerging across the Fortune 500. Yet even the most well-managed organizations face a common challenge: how to align numerous AI initiatives toward a cohesive strategic vision.

When your organization has dozens or even hundreds of AI projects running simultaneously, tactical excellence alone isn’t enough. Each team naturally optimizes for their specific objectives—whether that’s customer experience, operational efficiency, or product innovation. What’s often missing is a clear North Star that unifies these efforts.

This is where a Corporate AI Governance Manifesto becomes invaluable. Read more below. -s

https://t.co/NeIU8NTXmR