Chris Sparshott

Always building and learning.

[Chris Sparshott @seemly]

Some ideas on how you could turn boring businesses into your very own business.

Obviously there's much more you could do on top of this!

[Chris Sparshott @seemly]

@staticmaker1 @KryptoKomplex Create a database of boring businesses, and categorise them by type, and tag them by feature.

Don't just publish newsletters. Create them as blog posts, containing a more detailed write up, with more details (if possible).

With a large enough database you can create comparisons

[Chris Sparshott @seemly]

@aleyda More tweets than happy customers.

[Chris Sparshott @seemly]

@TheGriftReport @dmpcarey @TheNewsAgents How rich are we talking? £5m per person "gifts" that aren't taxable, and undeclared, rich?

If a socialist MP had received a £5m gift, I wonder if you'd be so blasé and forgiving?

I want MP's to be untarnished, with no fear of them being buyable.

[Chris Sparshott @seemly]

@milto28796 @PoliticsJOE_UK @Nigel_Farage @reformparty_uk @GBNEWS @GoodwinMJ @afneil @megynkelly Can you share your proof of singular gifts of £5m, please?

[Chris Sparshott @seemly]

@aleyda If I was one of the human engineers, my day rate would be astronomical.

[Chris Sparshott @seemly]

No guarantees about the factuality of that real-time information, though. Good ol' American freedom of speech!

"Agents now have access to the best real-time information source in the world."

[Developers @XDevelopers]

Announcing the hosted X MCP.

Agents now have access to the best real-time information source in the world.

Connect Grok, Cursor, or any MCP-compatible AI tool to the X API without any setup!

Check it out here: https://t.co/5MzPYwGFzD

[Chris Sparshott @seemly]

@nicksdot It seems that the US is becoming the enemy of the world in almost every way.

There must be a better way for humans to move forward.

[Gareth Wild @GarethWild]

Accurate 🔥🔥🔥🔊🔊🔊

[Chris Sparshott @seemly]

Why are these demos always of fighting or dancing robots?

We want robots to do mundane tasks such as dishes, laundry, ironing, cooking, housework, etc.

[Eren Chen @ErenChenAI]

Feel the intimidation factor of this 1.82m humanoid robot.

[Andy Chadwick @digitalquokka]

I asked ChatGPT for a running shoe recommendation twice last week. Same question, same model. I changed one sentence about who I was.

As a CrossFitter training for a marathon: Adidas Adizero Evo SL.

As a rock climber training for the London Marathon: ASICS Novablast 5.

Both lists included a Brooks Ghost. Different versions. The CrossFitter got the 18, the climber got the 17.

A new study just proved this at scale. 2,000 runs across OpenAI and Anthropic models. When the buyer persona changed, mid-market brands saw up to 75% of their recommendations replaced. Household names barely moved.

Which means your brand doesn’t have one AI ranking. It has one per customer type.

And it means most LLM prompt tracking is measuring the wrong thing. Your tool probably reruns prompts daily, which helps. But one run a day is one sample (identical prompts only kept about half the list between runs), the dashboard wiggles get read as signal, and the prompts are still generic: a customer who doesn’t exist.

The fix: build prompts around your real customer segments, repeat the runs, and report visibility per persona. Can’t afford the full matrix? Pick your single most valuable persona and put it on the front of everything. Knowing exactly who your customer is was always the job. The LLMs have just started punishing brands that never decided.

Full guide here 👉 https://t.co/ha3Isxq14m

Who does your brand show up for? Do you actually know?

[Chris Sparshott @seemly]

@johnnyratcliffe Looks like Spotify:

- Axed their design team due to AI adoption.
- Asked AI for marketing ideas.

[Chris Sparshott @seemly]

@PolitiKeyUK @danbarker > "It is highly likely that the user nudged or explicitly biased the Al to get this result."

Not unlike the starting of your prompt: "An idiot shared this..."

What a bellend you are.

[Chris Sparshott @seemly]

@JosephStash @danbarker That wasn't the issue raised by the OP though.

He was looking for food, and complained nowhere was open. Dan proved otherwise.

Also, London is ~2x larger in size than New York, but both have a similar population.

[DeepakNess @DeepakNesss]

Occasionally I still visit Bill Slawski's blog "SEO by the Sea", and sometimes it's either down or slow. So I created a searchable archive of the entire website in an open-source GitHub repo.

Other SEO folks may also find it helpful: https://t.co/8vrNLJeC4l

[Chris Sparshott @seemly]

@danbarker Maybe announcing/publishing how the warehouse was found wasn't the brightest move.

They just made future "investigations" harder...

[Paulund @paulund]

Can we agree a standard on what folder skills go in.

I see too many repos with .Claude and .github and .opencode all with duplicate skills in.

Can't we just have .agents folder?!

[Chris Sparshott @seemly]

@VicVijayakumar It's just the times we live in!

We're still young at heart!

[Chris Sparshott @seemly]

@JasonBanks67435 @ToddBonzalez5 @TheJoeySwoll Almost doing something means you did absolutely nothing.

I almost won the lottery jackpot, but won nothing.

Bull Shark Mentality.

[Chris Sparshott @seemly]

@VicVijayakumar GlassDoor would be a better suitor, as I tend to look at company reviews before applying for jobs as part of due diligence.

[Chris Sparshott @seemly]

@VicVijayakumar You can invite others to openly network with you for this purpose?

You're extremely limited in number of people you can message on LinkedIn outside your network (unless Premium).