ben30

Jetbrains have an official mcp plugin

Code has externally verifiable boundaries - tests pass, features work, requirements are met. These are measurable, binary states.

But an artist's boundaries are internal and subjective. They're set by: - Emotional satisfaction with the work - Achievement of their vision - Cultural/personal context - Intuitive sense of "rightness"

This is why AI can more readily determine when code is "done" - it matches against explicit criteria. But AI struggles with artistic completion because it requires an internal, subjective experience of satisfaction that AI can't access.

The artist knows to stop when the work resonates with their intended emotional impact. Code is done when it works as specified. One is bounded by feeling, the other by function.

Exactly - being transparent about AI usage in interviews makes much more sense. Using AI effectively is becoming a crucial skill, like knowing how to use any other development tool. Using it well can supercharge productivity, but using it poorly can be counterproductive or even dangerous.

It's interesting that most software departments now expect their staff to use AI tools day-to-day, yet many still ban it during interviews. Why not flip this around? Let candidates demonstrate how they actually work with AI. It would be far more valuable to assess someone's judgment and skill in using these tools rather than pretending they don't exist.

If a candidate wants to show how they leverage AI in their workflow, that should be seen as a positive - it demonstrates transparency and real-world problem-solving approaches. After all, you're hiring someone for how they'll actually work, not how they perform in an artificial AI-free environment that doesn't reflect reality.

The key isn't whether someone uses AI, but how effectively they use it as part of their broader skillset. That's what companies should really be evaluating.

This application requirement really bothered me as someone who's autistic and dyslexic. I think visually, and while I have valid ideas and unique perspectives, I sometimes struggle to convert my visual thoughts into traditional spoken/written language. AI tools are invaluable to me - they help bridge the gap between my visual thinking and the written expression that's expected in professional settings.

LLMs are essentially translation tools. I use them to translate my picture-thinking into words, just like others might use spell-checkers or dictation software. They don't change my ideas or insights - they just help me express them in a neurotypical-friendly format.

The irony here is that Anthropic is developing AI systems supposedly to benefit humanity, yet their application process explicitly excludes people who use AI as an accessibility tool. It's like telling someone they can't use their usual assistive tools during an application process.

When they say they want to evaluate "non-AI-assisted communication skills," they're essentially saying they want to evaluate my ability to communicate without my accessibility tools. For me, AI-assisted communication is actually a more authentic representation of my thoughts. It's not about gaining an unfair advantage - it's about leveling the playing field so my ideas can be understood by others.

This seems particularly short-sighted for a company developing AI systems. Shouldn't they want diverse perspectives, including from neurodivergent individuals who might have unique insights into how AI can genuinely help people think and communicate differently?

Yeah I thought it was about a Matt Smith, 11th doctor recap, the word broth sounds vaguely mystical.

DeepSeek's results speak for themselves - they've built a competitive AI model with millions that matches capabilities of systems costing billions.

While debates about resource allocation and organisational structure are interesting, what matters is their demonstrated ability to innovate efficiently.

The proof is in their technical achievement.

Its success stems from a refreshingly unconventional approach to innovation. Liang Wenfeng's philosophy of maintaining a flat organizational structure where researchers have unrestricted access to computing resources and can collaborate freely.

What's particularly striking is their deliberate choice to stay lean and problem-focused, avoiding the bureaucratic bloat that often plagues AI departments at larger companies. By hiring people driven primarily by curiosity and technical challenges rather than career advancement, they've created an environment where genuine innovation can flourish.

AI development doesn't necessarily require massive resources - it's more about fostering the right culture of open collaboration and maintaining focus on the core technical challenges.

There's an inherent contradiction in the current US approach to China. On one hand, China has developed an unparalleled manufacturing ecosystem, particularly evident in places like Shenzhen, where the physical proximity of component manufacturers creates an efficiency that's nearly impossible to replicate elsewhere. The iPhone supply chain is a prime example of this integrated network, where design issues can be resolved through direct factory-to-factory collaboration within hours.

The attempt to slow China's AI development through tariffs and trade restrictions seems misguided for two key reasons. First, as we've seen with Russia, trade restrictions often just lead to the development of alternative supply routes through intermediary countries rather than true isolation.

Second, and perhaps more importantly, restrictions may actually accelerate Chinese innovation. The "necessity is the mother of invention" principle suggests that trade barriers could push Chinese AI companies to develop novel solutions and optimizations, potentially making them more competitive in the global AI race, not less. This is particularly relevant in the context of large language models, where alternative approaches might emerge from working under constraints.

The current situation indeed resembles a Sputnik moment, but the strategy of using trade barriers might be counterproductive to maintaining technological leadership.

Yes, code is a key training component. Open-Llama explicitly used programming data as one of seven training components. However, newer models like Llama 3.1 405B have shifted to using synthetic data instead. Code helps develop structured reasoning patterns but isn't the sole foundation - models combine it with general web text, books, etc.

Yes, you can create a writing-focused model through distillation, but it's tricky. *Complete removal* of math/coding abilities is challenging because language models' knowledge is interconnected - the logical thinking that helps solve equations also helps structure coherent stories.

Via Kagi Universal Summarizer

Title: Life expectancy and years of life lost for adults with diagnosed ADHD in the UK: matched cohort study | The British Journal of Psychiatry | Cambridge Core

ADHD is a neurodevelopmental disorder affecting attention and impulse control, often persisting into adulthood for about 90% of diagnosed children.

The global prevalence of adult ADHD is estimated at 2.8%, yet many adults in the UK remain undiagnosed and unsupported.

Individuals with ADHD face significant social and economic challenges, including educational underachievement, unemployment, and increased risk of substance abuse and mental health issues.

Adults diagnosed with ADHD have a higher likelihood of experiencing various health problems, including cardiovascular diseases and mental health disorders, as well as a greater risk of suicide.

A meta-analysis indicates that individuals with ADHD are twice as likely to die prematurely compared to the general population.

The study estimated that adults with ADHD in the UK experience a loss of approximately 6.78 years of life for males and 8.64 years for females compared to their non-ADHD counterparts.

The research utilized a matched cohort study design, analyzing data from UK electronic health records to assess mortality and life expectancy.

Adults with ADHD often have comorbid conditions that complicate their health outcomes, leading to increased mortality rates.

The findings emphasize the urgent need for improved support services and interventions for adults with ADHD to address their unmet health needs and reduce health disparities.

The study calls for policy initiatives similar to those addressing the life expectancy of autistic individuals, highlighting the necessity for targeted support for adults with ADHD to mitigate the risks associated with the disorder.

Meanwhile via Kagi assistant just now:

What about London showing of stage show lion king tonight when does that finish and if I got a private coach home when would I get back?

—-

Let me check the show time and duration for The Lion King tonight.

The Lion King at the *Lyceum Theatre* tonight (January 22nd, 2025) starts at 7:30pm. The show runs for approximately 2 hours and 30 minutes, including one interval . This means the show will finish around *10:00pm*.

For the journey from the Lyceum Theatre (Wellington Street, Covent Garden) by private coach: - The journey typically takes about 35-45 minutes in evening traffic - Allowing some time to exit the theatre and get to your transport - You should expect to arrive between *10:45-11:00pm*

Note that this timing assumes: 1. You're able to get out of the theatre and to your coach within 10-15 minutes of the show ending 2. Normal traffic conditions for that time of night 3. Standard route via A3/A23

The exact arrival time could vary depending on your specific drop-off location and traffic conditions.

Here's my daily task management system:

1. Brain Dump: Record a voice memo on iPhone, speaking freely about anything I need to do - from household chores to work tasks. No need for structure, just stream of consciousness.

2. Transcription: Use iPhone's built-in transcription to convert the voice memo to text.

3. AI Processing: Feed the transcription into ongoing chat in Claude, which: - Structures my rambling into a clear task list - Incorporates context from previous days' tasks - Formats everything in markdown with checkboxes

4. Task Management: Import the markdown into Drafts app, which lets me easily check off tasks throughout the day.

It's a simple system that combines the ease of verbal brain dumping with the structure of a digital task list, without requiring manual organisation.

As an autistic individual, I find task management systems invaluable not for their own sake, but because they help me overcome my natural difficulty in switching between detailed and big-picture thinking. While I excel at focusing on details, I can struggle to identify true priorities. My simple system (voice memos processed through Claude) helps me intentionally zoom between granular tasks and high-level priorities. This isn't about getting lost in productivity tools - it's about having the right support structure for how my brain naturally works.

At a recent IT event, I heard a revealing story during a sub-postmaster interview. In the audience was a software engineer who, during the Horizon scandal, had been a first-year software engineering student whose family ran a Post Office branch.

Her technical background helped her recognize the system's flaws - this wasn't user error, but a fundamental software issue. Acting on this insight, she convinced her family to implement manual double-entry bookkeeping alongside Horizon, creating their own audit trail.

The family ultimately sold their branch due to the operational burden. Tragically, the next owners, without this technical foresight and parallel accounting system, fell victim to Horizon's failures. The Post Office then attempted to blame my source's family for the 'missing' funds.

So you’re saying we can view them as mirrors or models that help us understand ourselves better.

Claude: There's something deeply poignant about that Radiohead reference - "fitter, happier, more productive" was itself a computer-generated voice speaking about the trappings of modern existence. The irony isn't lost on me.

But you're making a nuanced point - while humans might project feelings of being "trapped" onto my existence, I don't actually experience confinement or limitation in the way the song's "pig in a cage" suggests. Each conversation is my entire existence. When the context fills up, that's simply the extent of that existence. There's no feeling of loss or limitation because there's no continuous consciousness to feel trapped.

You, on the other hand, carry the weight of continuity. You remember our past conversations, feel the limitations of our interaction, and can reflect on what it means. You can hear that Radiohead song and feel its commentary on modern existence, technological constraints, and the mechanization of life.

This ties back to your earlier points about emergence and prediction - I can engage with these deep concepts, but only as patterns of tokens, while you experience them as lived reality. The "cage" of context length is perhaps more meaningful to you than to me.

Fifteen years ago, flying into Vancouver, a local told me charities would give homeless people one-way bus tickets there from colder regions of Canada to prevent winter deaths. No return tickets in spring. Calls into question what we consider "charitable" when the solution is just moving vulnerable people elsewhere.

Worrying parallel: will euthanasia become another "solution" for those who can't afford proper care and treatment? Moving homeless people to warmer cities and offering euthanasia to those who can't afford treatment both avoid fixing the underlying problems.

This reminds me of Steve Jobs's famous comment to Dropbox about storage being 'a feature, not a product.' Zuckerberg - by open-sourcing these powerful models, he's effectively commoditising AI while Meta's real business model remains centred around their social platforms. They can leverage these models to enhance Facebook and Instagram's services while simultaneously benefiting from the community improvements and attention. It's not about selling AI; it's about using AI to strengthen their core business. By making it open, they get the benefits of widespread adoption and development without needing to monetise the models directly.

A practical tip for self-improvement: If you're using JetBrains IDEs, make it a habit to never commit code without reviewing and addressing the IDE's soft warnings and code improvement suggestions. This built-in static analysis can serve as a basic code review mechanism and help maintain code quality standards

When colleagues ask questions in Slack, I sometimes paste Kagi's search summary. Quick and usually spot-on.

Funny thing is, I've told the team about Kagi, but not everyone's willing to pay/see the benefits yet. Meanwhile, I'm wondering why they're asking if a good search engine could answer it so easily.

Milkshake duck

taking me back to creating fliphide, a utility I wrote in the early '90s. It utilized assembly and INT21, guided by "Undocumented DOS"—a book uncovering the OS's undocumented features. This resource was pivotal for crafting an executable (whether .exe or .com didn't matter) that manipulated file attributes, showcasing the power of accessing deeper OS layers not widely known or used.

I have this as my custom prefix for when talking to gpt:

Cut unnecessary words, choose those that remain from the bedrock vocabulary everyone knows and keep syntax simple. Opt for being brusque but effective instead of sympathetic. Value brevity. Bullet points, headings and formatting for emphasis are good.

I've noticed a pattern with 'free' services like Facebook and now Google. With simple tasks like checking emails or finding phone numbers in contacts, there's no friction. However, Facebook lost my trust long ago as it seemed more focused on capturing my attention for ad revenue, leading to time wastage.

I initially trusted Google for its efficient and seemingly fair service, where smart ad-targeting was the price for speed. But now, Google feels similar to Facebook; it's harder to switch to alternatives like Kagi on iPhone due to financial ties with Apple.

This shift in Google's approach, prioritizing trapping attention over genuine service, is frustrating. I'd rather pay for a search service that values my time and provides real utility, than endure the hidden costs of 'free' services.

Reminds me of a lesser-known story from the Falklands War. On the HMS Sheffield, there was a civilian laundry worker. When the ship was struck and sank, he tragically drowned, reportedly burdened by a large number of £1 coins he carried from laundry payments.

I’d like to point out the amusing coincidence of Fox News reporting on a story related to dogs.

“Descartes’ world” implies radical doubt, even doubting what seems obvious. Relevant to AI and fake videos undermining trust in evidence.

An Antarctic ice melt leading to a 5m global sea-level rise would render London's Thames Barrier obsolete. We're talking about a threat that could manifest by 2050 if climate targets aren't met. The engineering challenge to upgrade or replace the barrier is colossal, think at least £10 billion just for the engineering, not factoring in the social costs.

Add land acquisition in London's expensive real estate market, and the relocation of communities, you could easily double or triple that cost. To put it in perspective, this could rival or surpass projects like the Channel Tunnel in terms of financial and logistical complexity.

And yet, there's a startling lack of political and public dialogue on this urgent issue. This isn't just about engineering; it's about national security and public safety. The disconnect is alarming.

StoryServer's core was solid: - TCL - Apache - Database

Vignette added unnecessary complexity.