Alphero Intelligence

One of my latest learnings on my accessibility journey was discovering that deaf and Deaf mean different things.

deaf (lowercase d) refers to the medical condition of hearing loss, regardless of sign language use, cultural identity, or when it occurred.

Deaf (capital D) describes a cultural and linguistic identity centred around sign language, such as NZSL.

It was a simple distinction, but one that reminded me how much language matters when we're designing for inclusion.

Woah, this is cool! WikiCity transforms 100,000 most-read Wikipedia articles into a 3D city. The tallest skyscrapers represent the most popular topics. It's a fascinating example of how creative data visualisation can turn familiar information into something completely new to explore. https://wikicity.app

One of my latest learnings on my accessibility journey was discovering that deaf and Deaf mean different things.

deaf (lowercase d) refers to the medical condition of hearing loss, regardless of sign language use, cultural identity, or when it occurred.

Deaf (capital D) describes a cultural and linguistic identity centred around sign language, such as NZSL.

It was a simple distinction, but one that reminded me how much language matters when we're designing for inclusion.

Woah, this is cool! WikiCity transforms 100,000 most-read Wikipedia articles into a 3D city. The tallest skyscrapers represent the most popular topics. It's a fascinating example of how creative data visualisation can turn familiar information into something completely new to explore. https://wikicity.app

One of my latest learnings on my accessibility journey was discovering that deaf and Deaf mean different things.

deaf (lowercase d) refers to the medical condition of hearing loss, regardless of sign language use, cultural identity, or when it occurred.

Deaf (capital D) describes a cultural and linguistic identity centred around sign language, such as NZSL.

It was a simple distinction, but one that reminded me how much language matters when we're designing for inclusion.

Woah, this is cool! WikiCity transforms 100,000 most-read Wikipedia articles into a 3D city. The tallest skyscrapers represent the most popular topics. It's a fascinating example of how creative data visualisation can turn familiar information into something completely new to explore. https://wikicity.app

Google Cloud has scaled its gen-AI use case list from 101 to 1,001 in 18 months, showing just how quickly AI adoption is accelerating. Organisations are rapidly shifting from experimentation to measurable impact. Lots of inspiration here for what’s possible. Explore the full list.

I should know better than to be surprised by scammers, but honestly, sometimes their creativity deserves a slow clap.

The latest one? A fake meeting invite claiming there’s been an unexpected McAfee charge. The email went to spam, but the invite still snuck into my calendar. Sneaky!

Stay sharp out there. If it smells fishy, it probably is (and not in a good way).

The current architecture of LLMs means that prompt injection is a game of whack-a-mole and may never be entirely solvable. We will likely see a continuous race between guardrails and new exploit techniques for some time.

This non-technical overview of the problem sums it up nicely

There is not a lot of technical information available on using DeepSeek r1 on Perplexity. But on the surface of things, this looks like a slightly better way to access the DeepSeek r1 model (if you can't run it locally yourself). As the DeepSeek model is run on US and EU servers without sending data off to China.

It is still an orange status, as they still share certain information with US companies and partners.

With all the noise around AI, it's great to see practical real-world examples of how AI can help humanity with some of its really complex problems.

Take a look at how AI has almost done it

Drunken animals, bottles, a gun, an actress... look closely enough and you might spot something unexpected in the drip-laden layers of Pollock's masterpieces.

Pure creative genius or coincidence? Take a look for yourself.

Google Cloud has scaled its gen-AI use case list from 101 to 1,001 in 18 months, showing just how quickly AI adoption is accelerating. Organisations are rapidly shifting from experimentation to measurable impact. Lots of inspiration here for what’s possible. Explore the full list.

The current architecture of LLMs means that prompt injection is a game of whack-a-mole and may never be entirely solvable. We will likely see a continuous race between guardrails and new exploit techniques for some time.

This non-technical overview of the problem sums it up nicely

With all the noise around AI, it's great to see practical real-world examples of how AI can help humanity with some of its really complex problems.

Take a look at how AI has almost done it

Drunken animals, bottles, a gun, an actress... look closely enough and you might spot something unexpected in the drip-laden layers of Pollock's masterpieces.

Pure creative genius or coincidence? Take a look for yourself.

I should know better than to be surprised by scammers, but honestly, sometimes their creativity deserves a slow clap.

The latest one? A fake meeting invite claiming there’s been an unexpected McAfee charge. The email went to spam, but the invite still snuck into my calendar. Sneaky!

Stay sharp out there. If it smells fishy, it probably is (and not in a good way).

There is not a lot of technical information available on using DeepSeek r1 on Perplexity. But on the surface of things, this looks like a slightly better way to access the DeepSeek r1 model (if you can't run it locally yourself). As the DeepSeek model is run on US and EU servers without sending data off to China.

It is still an orange status, as they still share certain information with US companies and partners.

Google Cloud has scaled its gen-AI use case list from 101 to 1,001 in 18 months, showing just how quickly AI adoption is accelerating. Organisations are rapidly shifting from experimentation to measurable impact. Lots of inspiration here for what’s possible. Explore the full list.

There is not a lot of technical information available on using DeepSeek r1 on Perplexity. But on the surface of things, this looks like a slightly better way to access the DeepSeek r1 model (if you can't run it locally yourself). As the DeepSeek model is run on US and EU servers without sending data off to China.

It is still an orange status, as they still share certain information with US companies and partners.

Drunken animals, bottles, a gun, an actress... look closely enough and you might spot something unexpected in the drip-laden layers of Pollock's masterpieces.

Pure creative genius or coincidence? Take a look for yourself.

I should know better than to be surprised by scammers, but honestly, sometimes their creativity deserves a slow clap.

The latest one? A fake meeting invite claiming there’s been an unexpected McAfee charge. The email went to spam, but the invite still snuck into my calendar. Sneaky!

Stay sharp out there. If it smells fishy, it probably is (and not in a good way).

The current architecture of LLMs means that prompt injection is a game of whack-a-mole and may never be entirely solvable. We will likely see a continuous race between guardrails and new exploit techniques for some time.

This non-technical overview of the problem sums it up nicely

With all the noise around AI, it's great to see practical real-world examples of how AI can help humanity with some of its really complex problems.

Take a look at how AI has almost done it