Matt Stringham

15 Years Digital Marketing, Web Development. I test AI tools for Marketing so you don't have too. 🦞🪽

[Matt Stringham @Stringhammer86]

@AlexFinn Open weights are going to dominate especially if government keeps getting involved.

[Matt Stringham @Stringhammer86]

The potential is endless.

For us…

✅ Media buys
✅ Conversion tracking and attribution
✅ API integrations
✅ Reporting
✅ Copywriting
✅ A/B tests (automated)
✅ ILM and Data extraction/verification
✅ OCR

[Tibo @thsottiaux]

ChatGPT Work is for

✅ Creating and hosting sites
✅ Managing your emails for you
✅ Summarizing mountains of documents
✅ Creating top-notch docs, sheet and slides

Already in your mobile app or on https://t.co/WtXdTqu9lT, included in Plus, Pro, Business and Enterprise plans

[Matt Stringham @Stringhammer86]

Washington didn’t stop AI progress. It BOOSTED Open-weight.

Good for Kimi bad for frontier models.

Policy friction has a price.

June 9: Fable 5 launch
• June 12: government-ordered suspension
• June 16: GLM‑5.2 open-weight release
• June 26: GPT‑5.6 broad rollout delayed
• July 1: Anthropic restrictions lifted
• July 9: GPT‑5.6 public release
• July 16–17: Kimi K3 release and benchmark results

[Jun Song @jun_song]

The massive capital expenditure mentioned here is mostly used as CAPEX, with only about 20%~ used for actual training.

This is because US companies spend 5x more on serving just to keep things closed, as they have to directly serve a massive consumer base.

If you release it as open-weight, there is no need for this direct serving, which eliminates that astronomical spending.

That is exactly what the chart below explains.

This is not a dumping, it’s a strategy.

[Matt Stringham @Stringhammer86]

The balance is critical. One hold back and open weight caught up.

[Dean W. Ball @deanwball]

In a matter of weeks, U.S. federal AI policy has gone from implausibly libertarian to increasingly draconian and opaque. Today, over 35 distinct observations, I analyze how we got here and offer the most succinct statement I can about what exactly I propose we should do next.

[Matt Stringham @Stringhammer86]

Washington didn’t stop AI progress. It constrained who could ship it.

While Fable 5 was suspended and GPT-5.6 delayed, Chinese/open-weight challengers advanced. Weeks later, Kimi K3 was ~3 points from the U.S. frontier—and ahead on frontend coding.

Policy friction has a price.

June 9: Fable 5 launch
• June 12: government-ordered suspension
• June 16: GLM‑5.2 open-weight release
• June 26: GPT‑5.6 broad rollout delayed
• July 1: Anthropic restrictions lifted
• July 9: GPT‑5.6 public release
• July 16–17: Kimi K3 release and benchmark results

[Matt Stringham @Stringhammer86]

I can confirm this. Terra is publishing media buys and ads at 2x

[Peter Steinberger 🦞 @steipete]

5.6 Terra high is underrated. Switched @clawsweeper (GitHub review bot) to it and it's ~40% faster overall with negligible quality loss. Better than 5.5 on all counts. Massively cheaper.

(Tried xhigh but that negates perf wins, didn't make a noticable difference in review evals)

[Matt Stringham @Stringhammer86]

#AI has completely leveled the playing field. The only jobs it'll take are those who aren't willing to grab the opportunity.

[Matt Stringham @Stringhammer86]

The real AI disruption isn’t just better benchmarks.

It’s collapsing the cost of intelligence.

Kimi (Moonshot AI)—an open model—is now competitive with the leading closed models across several coding benchmarks. In the results shown here, it leads on Program Bench and SWE Marathon, while nearly matching GPT-5.6 on Terminal Bench 2.1.

That matters far beyond software development.

For marketing teams, the long-term opportunity is to move high-volume, repeatable work away from metered API calls and into privately operated AI infrastructure:

• Content research and first drafts
• SEO clustering and content refreshes
• Campaign QA and reporting
• Lead enrichment and classification
• Customer-review analysis
• Knowledge-base search
• Creative variation and personalization

Today, running a model at this level locally still requires hardware most businesses won’t own—a multi-GPU system or substantial unified-memory machines. So this is not the end of ChatGPT, Claude, or paid APIs.

It is the beginning of a hybrid AI economy.

Use frontier cloud models where maximum capability matters. Route predictable, token-heavy workloads to smaller or open models where the economics, privacy, and control are better.

At scale, the savings aren’t about avoiding a $20 subscription. They come from reducing millions of recurring input and output tokens across automated workflows—without paying a margin on every inference forever.

The important signal is the trajectory. Local AI is becoming more capable and compute-efficient at a remarkable pace. As hardware improves and models become smaller, the break-even point will keep moving downmarket.

The marketing leaders who win won’t simply “use AI.” They’ll build intelligent routing systems that assign each task to the right model based on quality, speed, privacy, and cost.

Local AI has arrived. The next competitive advantage is learning where it belongs in the operating model.

#ArtificialIntelligence #MarketingOperations #MarketingAI #GenerativeAI #DigitalTransformation

[Kimi.ai @Kimi_Moonshot]

Introducing Kimi K3: Open Frontier Intelligence

🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows

Kimi K3 is now live on on https://t.co/zrk6zZxZUo, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.

🔗 API: https://t.co/XCrgjXAqMw
🔗 Tech blog: https://t.co/YTfiMSNM1f

[Matt Stringham @Stringhammer86]

@DaveShapi Savant Syndrome for sure. I go from 'wow' and completely blown away to "wtf" pretty quick at times.

[Matt Stringham @Stringhammer86]

If you're not running both @openclaw and @NousResearch (Hermes) I don't know if you're truly winning.

They're both smart, both capable but both have their issues but getting better and better.

Openclaw understands Ops really well and feels native to talk to.

Hermes is self learning and gets better with time.

BOTH help fix each other. Hermes for 🦞 updates and Openclaw for 🪽 clarifications when it gets stuck and its over my head (OCR, ILM, etc.)

I'm convinced not one product is always going to be your absolute go-to. Everyone will have a tech stack that supports their workflows.

#AI

[Matt Stringham @Stringhammer86]

@valuedrift This is my hope. OpenAI stake.

[Matt Stringham @Stringhammer86]

@RyDawg42 @aleabitoreddit Same.

[Matt Stringham @Stringhammer86]

@Craigt8484 @RyDawg42 This. And Yeah they didn’t disclose spcx either. I would’ve bailed. Everyone saw that coming you’d think as a fund they’d have better edging.

[Matt Stringham @Stringhammer86]

#AI doesn’t replace the marketer.
It gives the marketer more leverage.

This week, we used a human-led, AI-enabled process to turn a fragmented ecommerce product catalog into a clean paid-social launch system:

• 19 product ads structured
• 96 product images mapped and QA’d
• 19 primary-text variations updated
• Every ad packaged for clear client approval
• Nothing launched without human sign-off

The humans set the strategy, exclusions, voice, approvals, and launch rules.

AI handled the repeatable execution: organizing assets, mapping products to creative, applying approved revisions, checking consistency, and assembling the review layer. UTMs etc.

That’s the real opportunity: not “AI marketing” with no accountability.

Human judgment + AI leverage + a clear approval process.

How is AI helping your agency?

Tech stack:
🪽@NousResearch
🦞 @openclaw
@OpenAI / @Meta

[Matt Stringham @Stringhammer86]

@colinsolvely @MichaelGannotti Openclaw is smart out of the gate. The challenges in its updates are worth it IMO.

Hermes has great potential but it needs a lot of guidance. My 🦞 struggles with updates but once running its performance is solid.

[Matt Stringham @Stringhammer86]

Whoa.

[Ruben @rdominguezibar]

Demis Hassabis just published the most consequential essay in AI this year. Nobel Prize winner. CEO of Google DeepMind. The person who built AlphaFold.

His estimate: AGI in a few years. His comparison: not the internet. Electricity. Fire.

The number that should stop you cold is this one. 10x the Industrial Revolution at 10x the speed. The Industrial Revolution took 100 years to reshape civilization. He is describing something an order of magnitude larger happening in a decade.

His policy proposal is the part nobody is covering:

▫️ A FINRA-style Standards Body for frontier AI
▫️ Labs submit models 30 days before release for independent testing
▫️ Benchmarks updated quarterly, built independently from labs
▫️ The body could coordinate a slowdown among all Frontier Labs if needed

That last bullet is the one. The CEO of Google DeepMind just publicly proposed a legal mechanism to pause AI development if the situation demands it.

He ended with this: "We've essentially found a way to make sand think."

Either the most important sentence written this decade or the most embarrassing. No in between.

[Matt Stringham @Stringhammer86]

AI agents are moving beyond chatbots.

We just completed an enterprise pilot where AI analyzed a large archive of poorly labeled historical records, compared conflicting metadata sources, identified cataloging errors, and generated evidence-backed corrections.

But the valuable part wasn’t simply “using AI.”

We built a controlled workflow around it:

• Validate source files before trusting them
• Compare old and new metadata systematically
• Sample the underlying records for evidence
• Flag uncertain cases for human review
• Map AI output to an approved vocabulary
• Measure accuracy before scaling

The pilot worked.

Now we’re moving into the next phase: connecting the system to approved APIs and testing it as a repeatable program.

This is where enterprise AI gets interesting.

Not replacing experts.

Giving them a system that can investigate thousands of records, surface the exceptions, and let humans focus on the decisions that actually require judgment.

The companies that win with AI won’t just have the best model.

They’ll build the best validation loop around it.

#ai @OpenAI @openclaw @GeminiApp

[Matt Stringham @Stringhammer86]

the BEST value for the $$$ right now.

[Sam Altman @sama]

GPT-5.6 sol is half the price and ~twice as token efficient as fable in many cases for accomplishing the same task.

happy to deliver at one-quarter of the price.

[Matt Stringham @Stringhammer86]

@AlexFinn It'll probably be faster than Fable 5.

My last administrative task I had to ask it where it went 😅

[Matt Stringham @Stringhammer86]

This is the only thing holding back the🦞. It 100% outperforms everything else that I use, but as soon as I update it or make a change somewhere, it's almost devastating.

But I think it's also making me an open-claw professional at this point.
#openclaw