Matt Brighton (@mattbrighton_) 's Twitter Profile
Matt Brighton

@mattbrighton_

A product manager who can code. Also make YouTube videos

📹 youtube.com/mattbrighton
🏡 propertyinsights.io
📲 stampdutycalculator.net

ID: 1280073630282592259

linkhttp://tinyurl.com/propertycourse calendar_today06-07-2020 09:38:37

627 Tweet

726 Takipçi

311 Takip Edilen

Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

Shipping something is bigger than shipping nothing. Great reminder to just start and perfect it in future. You’ll also learn loads from releasing and testing in the real world.

Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

It blows my mind 70% of lottery winners lose it all or end up worse off. studies indicate that 70% of lottery winners end up bankrupt, conservative estimates say that nearly a third of winners declare bankruptcy, a rate significantly higher than the general population.

Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

I really want Gemini CLI to be great. But it's terrible for now. Really hoping Gemini 3 and future investment will help it become top of the market. It's very slow, takes a long time to make changes and makes a lot of mistakes. Despite having a large context window which is

I really want Gemini CLI to be great. But it's terrible for now.  Really hoping Gemini 3 and future investment will help it become top of the market. 

It's very slow, takes a long time to make changes and makes a lot of mistakes. Despite having a large context window which is
Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

There’s a massive ego problem with handwritten versus AI coding. Hence why manual engineers call it AI slop. At enterprise level, sure. But for the average small business or website users do not care if your ai code took a slightly longer route. The future will be split

Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

This week Claude code has been another 50% worse. 5 hour rate limit seems to have been secretly reduced? Codex is behaving 50% better and much faster than last week.

Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

I disagree. Start creating. You'll figure it out along the way. Nobody is perfect at the start. 99% of people are stuck in consumer mode. Start creating .

Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

Experimenting with free trials on propertyinsights.io with 10 free credits. Added this onboarding flow with 2 questions: 1. What's your biggest challenge 2. What are you hoping to achieve on here the free input is validated by AI to prevent garbage answers. That way, free

Experimenting with free trials on propertyinsights.io with 10 free credits.

Added this onboarding flow with 2 questions:

1. What's your biggest challenge
2. What are you hoping to achieve on here

the free input is validated by AI to prevent garbage answers.

That way, free
Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

New Feature ✨Company Ownership Search ✨ You can now search 4.3 million properties in the UK owned by companies at propertyinsights.co.uk Find out the company's ultimate beneficiary owners by following the Companies House ownership chains.

New Feature ✨Company Ownership Search ✨

You can now search 4.3 million properties in the UK owned by companies at propertyinsights.co.uk

Find out the company's ultimate beneficiary owners by following the Companies House ownership chains.
Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

Bit the bullet and bought the .co.uk domain for propertyinsights.co.uk after using .io for ages. It was up for ransom at $600usd. Hopefully a good investment. I found a competitor who had a very similar name spin up in the past 2 months using .co.uk and far surpassed on local

Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

Does SEO schema make a difference? I’ve heard a lot about it. And might encourage rich snippets. Have added it onto the live site and will report back

Matt Brighton (@mattbrighton_) 's Twitter Profile Photo

Initial thoughts on Claude Sonnet 4.5 - Way faster model, thinks and outputs better than previous models - Seems to be missing a lot of fixes and key things I point out, isn't following instructions properly. - When it does actually fix or create what I need, it's to a high