ghshephard

I was working on the clearly mistaken impression that "Made by Anthropic" in the cursor store meant that the artifact had been signed by Anthropic. If that is not the case, then two party's need to get their act together:

  - Cursor needs to add signing if they want to claim something was "made" by another party.

  - Anthropic needs to stamp out any provider from using said, "Made by Anthropic" until they start adding signing to their process.   Trademark law is useful here.

Until that happens - it's almost like you have to delete any extension from the Cursor Store off of your Cursor instance - kind of a crapshoot what you are actually downloading, no?

For example - what guarantee do I have with something like: " This publisher has verified ownership of cursor.com" - when I'm looking at the WSL extension. Are there any guarantees it's not been replaced with some other extension?

As someone who has read 20-30 takes on this topic, and talked about it endlessly for the last 2+ weeks, It's an epically interesting read for me (a non-mathematician) - just introducing me to the topic "Mathematization" was sufficient to make the entire article interesting. (I also learned about a new mathematician - Grothendieck).

The article was interesting enough that I spent about an hour reading excerpts and discussing in detail all of it's nuance at the breakfast table this morning. Strongly encourage a detailed read - it was great.

The developers would love to spend their time coding and crafting elite and elegant systems - but management at these companies has other things that they want them to do. Endless meetings, Alignment, Training, Constant Interviews, Quarterly Project Planning, Performance review...

Each one of these distractions can blow the mental context of a developer for hours as well - so not only do they get very, very little time to actually "program" - but the time they do get, is now mostly trying to get back into "flow" , re-establish a full mental context of the project, etc... So - that 5% is likely a huge over-estimate of the amount of "productive" time a developer has, as a good chunk of that is just recovering from the bullshit that the company inflicts on them so they can go heads down and actually start doing "work" - realistically I bet if you watched your average developer at a big FAANG - they spend substantially less than 2% of their time actually heads down coding productively.

Did you read the section on "Relationship to other sequences?"

https://www.superlogical.com/rex/docs/build/program-status#r...

Based on the last 5+ years at working at FAANG adjacent companies - less than 5% of a SWEs time is dedicated to coding (Exceptions of course for crazy things like porting massive code bases to Rust...) - If anything - now that "coding has been solved" - it significantly reduces the need for body-shops, because now you require strategic product decision making - much harder skill to hire for/develop than yet another LeetCoder.

So - definitely agree with your general thesis.

For anyone who has spent a lot of time in Singapore (I lived there for 5 years, from 2012 - 2017) - the following snippet isn't doing justice:

> pervasive CCTV and face recognition.

I'm willing to wager there is no country on the planet that comes within an order of magnitude in terms of their awareness of precisely where any given resident is to within a few meters.

In fact - I think we would struggle to find a country within 2 orders of magnitude - the combination of incredibly high technology, super density, lack of much in the way of "countryside" (and the jungles/parks have lots of camera's near the entrances) and a general societal consensus that omnipresent surveillance is worthwhile exchange for the world's safest country all come together to create a system in which you don't need to worry about obscuring who you are dating with/sleeping with/associating with.

The Government already knows, always has known.

I'll buy that - but, how often does one need an honest-to-goodness expert? 99% of the time most people's requirements can be satisfied with general-competence, and, while the AI systems we have today are mostly only generally competent in porting code from one language to another - they are getting consistently better. I expect we'll start to see general competence in many more areas in the next few years.

AI will likely replace divorce lawyers (or, at least make existing divorce lawyers much more efficient) - for non-contested divorces that are amicable. Roughly 95% of what a divorce lawyer does in that context is entirely automatable.

I guess in my case it makes sense because - I used the Apple Watch Ultra as ... a watch. The only two "non-watch" features that I really ever used were: - Find my Iphone/Make it beep. That stopped working once I disabled the ringer because it was annoyingly loud during meetings. - Setting a timer for 5 minutes or whatever during booking. Takes < 1 second to do that on my phone.

Everything else was stuff that the Casio does way better. And - need the phone for the camera - so it's always with me.

It's kind of amazing to have a watch with a 7+ year battery. I spent about 4 years with my Ultra. Might keep it for stuff like hiking where I want to track elevaton and such - but the other 99% of the time I expect the Casio will be my goto.

Just go type questions into the playground - leave "diffusion effect" on - you can see it for every question. https://chat.inceptionlabs.ai/

Absolutely 100% not true. I have colleagues in 4 "FAANG adjacent" companies plus the one I work at - zero of them have any faith in Enterprise Agreements from either OpenAI or Anthropic.

There's a reason why people spend more $$$ with Data Bricks, Palantir, AWS Bedrock etc.. and don't even consider using Anthropic or OpenAI APIs directly - it's because those guarantees provide very little in the way of data-discovery, audit requirements, or liquidated damages should it ever be discovered there was data leakage.

At least with these other companies, while the LD is likewise not great (typically limited to the amount of money you paid them) - you at least have some data-governance guarantees around running on dedicated hardware - no multi-tenancy, no third-party access outside of the AWS operators who keep the HW running - but are very much not in the business of looking at your data.

I think this is mostly a function of what's at risk - when company valuations get into the 10s of billions of dollars, the risk of IP leaking into what could be seen as competitive companies (OpenAI/Anthropic would be happy to take over the world - I don't sense that AWS or Azure, are as ruthless in stepping on their customers business, unless of course they are a SAAS provider) is just too significant a liability to take - particularly when you can de-risk.

All I can say with certainty that the legal team at our typical Multi-Billion dollar Silicon Valley company had zero faith in any licensing arrangements with Anthropic or OpenAI, regardless of they $$$ involved, and that even getting to the point where Amazon Bedrock on Dedicated GPUs (we're already a big AWS customer - so definite cost advantages to dealing with them) - took 4-6 months of legal review before we could allow our engineers to start using Claude and OpenAI coding agents. Still can't use Fable because of their Data Retention requirements.

If you are interested in the sandbox element, and space, and building - but want character - Satisfactory pulled me in for roughly 1200 hours over three years. I"ve got the scenery, and new discovery's burned in to my brain - elements of the landscape are more vivid in my memory than a lot of real places I've visited.

What's interesting is that there is zero procedural generation - single world with about 20 Biome, each one with all sorts of character. Monsters are incredibly well balanced, and the sheer terror I got from the spiders (which I think are a tad OP - especially the ones that launch)_ triggered my fight-flight reflex (mostly flight) more than any other game I can recall.

The game can be played in so many different ways too - you can take the factorio/big factory build out approach and speed run that, or just spend your time exploring, or, perhaps best of all - watch hundreds of hours of Youtube Videos to get a sense of the architect aesthetic that some of the top players bring to it. Epic Tech Tree. Cute back story.

I consider it the greatest Sandbox/space game ever - I don't know if we'll ever see anything like it again.

Correct thread - The Casio F-91W is the "FlipPhone" of watches. Going Retro on all our electronics. Retro Phone, Retro Watch, just walk away from the insane consumerism gadgetitis and find what makes you happy.

100% every word you wrote. I watched the Apple video. Thought it would be great to have this amazing GPS capable watch, with a cellular phone (that, I will confess, I did use three or four times and it was really good) - maps, remote finding of my phone, etc...

Honestly - if it wasn't for the incessant nagging me to upgrade, letting me know it's going to upgrade, letting me know it had failed to upgrade, and the never ending useless notification that I kept having to edit/unsubscribe/remove stupid apps that I never installed - I probably wouldn't have switched for a short period of time.

What I would have never predicted is how getting rid of all that noise has just made me so much happier.

Also - didn't realize how much I hate having to charge my watch every 16 hours.

I've been rocking out a Casio F-91W (Retails new for $23) for about 3 months now - And all the annoyance I had with with my Apple Watch Ultra (16 hour battery, constant annoyance to upgrade, endless and useless notifications) - is gone. The CR2016 battery last 5+ YEARS , replacements are < $2. Watch has everything I ever cared about - Date, Time, Alarms, Light, Stopwatch, water resistance.

And, I don't think I've once missed having the big annoying clunker of a watch.

If this is a "wake up call" - then your legal team needs immediate education.

First - there is this - https://openai.com/policies/how-your-data-is-used-to-improve... (linked from the Navier Stokes writeup)

I don't know how much more clearly they can write:

> When you use our services for individuals such as ChatGPT, Sora, or Operator, we may use your content to train our models.

One of the key selling tactics that companies like Data Bricks or Palantir provides their customers is "Data Governance" - that is, some control over where the data is being used. It's also a reason why enterprises don't use the OpenAI or Anthropic APIs directly - but through secondary sources that have Enterprise Agreements that do their best to make sure that no Company IP is ever retained by a third party, or even exists on a multi-tenant GPU. AWS Bedrock, and companies like together.ai, fireworks.ai have tons of deals that focus very much on data confidentiality.

The reality is - if you want any type of control - you run your own inference, on your own hardware. Anything else and you are at the mercy of third-parties, despite what their contracts might promise you.

Because they solved different problems, and where there was overlap, the solutions look different?

This feels like a concept that keeps getting rediscovered over and over. Feels like giving Agents (and their humans) a repository to throw ideas at (along with maturity, gates, context, implementation status, etc...) is going to be one of the next Billion dollar opportunity...

Another Blackboard Example: https://github.com/halbritt/striatum/tree/main/docs/rfcs

(Useful to see how this compares to another Blackboard type platform - Gastown https://github.com/halbritt/striatum/blob/main/docs/records/...)

Not just a model - It's a diffusion model. Instead of Next Token prediction it builds the entire page at once and then denoises it. Kind of mind blowing when you watch it happen.

Have you run into scenarios where a simple view doesn't accomplish what you require? I always feel like views do everything I want - because the speed is so great, the full recompute isn't that big a deal.

Maybe it's a bigger deal when you have multiple users/and or more repeated queries against something that's really expensive?

Holdonasec - I can list roughly (looking through my bookmarks and counting) - about 20sh different forum type sites with discussion (reddit, anand, verge, slashdot, ars, strachery, etc...) - None of them I like as much as HN. Why would trying to find a site that is better (given the OP didn't like HN) make me sound like a Bot?

And if the answer was, "Well, there is nothing better than HN" - kind of makes me wonder what his original point actually was.

Everyone gets $125 / month budget (compared to the $5k/month people get for coding LLMS) - it's pretty rare for anyone to get even close - maybe some of the administrative assistant who do nothing but type into CoWork all day go over - I think I use about $30-$50/month.

Our InfoSec teams are insanely restrictive. No Claude Code out of Sandboxes, MCP servers hyper locked down, etc.... I think your reservations are really good - but - at the end of the day - I need to do work - and CoWork does it for me.

It doesn't just have access to my entire machine - it has access to all of our Corporations Google Docs, Google Sheets, Jira Tickets, Confluence, Slack - everything (Except, weirdly - our GitRepos - they aren't confident enough there to let CoWork talk to them)

Weirdly - my work email is mostly irrelevant - nothing in their other than some sporadic company announcements - I'll go a week without looking at it - so it's the least concern. (Very different from personal gmail which, omg - if anyone got into - game over)

I do not want to confess how often I just drop in and change the order of my approximately dozen or so keys at the end of the ssh config file just to get the ones I want up to the top of the list.

So, I have this which I use about once a day:

  vis ()
  {
      vim ~/.ssh/config
  }

But - we have around 250k hosts in about 42 datacenters, some of them going back 8+ years, built by different teams - not everything has been standardized - and :

  $ wc -l ~/.ssh/config
  268 /home/shephard/.ssh/config

has lots and lots of weird IP mappings to ssh-keys that is just gross, and not something I would want to expose others to (mostly because it changes a lot behind the scenes)

So - from time to time, someone will say, "hey - the key that should be working in this datacenter/cluster isn't working, how do you get connected" - at which point I so a ssh -vvv and eyeball which one was accepted.

Being able to ask ssh just to print it out - would be awesome.

I don't think there has been a single month in the last 5 years that I haven't had to figure out which public-key was used to authenticate me to a host via `ssh -vv` - often to let someone else who is failing to connect know which key to use.

The `-Z` is less useful - as I almost always know exactly which and what order the public keys are - as most of our sshd instances fail after 5 attempts - so making sure either (A) the correct ones or used or (clumsily, B) - just putting the correct one in the first five to try.

Would be a nice feature to echo which key successfully worked.

Claude CoWork absolutely dominates - I use it continuously now when connecting to mail, calendar, jira, google docs - any type of knowledge work. Almost everyone I know (at work) who does "Knowledge work" type stuff (as opposed to coding) uses it as their primary pane-of-glass now.

I have 15+ years of top 1 percentile knowledge using awk to process complex unstructured documents - and I don't even bother anymore - just dump the dataset into CoWork and ask it to do the analysis, cross check its work.

I did it last night during a call for a reasonably complex 10,000 node cluster with all sorts of jobs/allocations/memory constraints. I didn't even attempt to throw together a quick awk script (which would have taken me 7-8 minutes - it was a beast) - just did a nomad-job-inspect of a couple hundred jobs, and dumped all the data at CoWork (real time, while talking on the zoom) and said, "Get me the Memory/CPU requirements based on job/datacenter/SKU constraints and give me a summary Table".

Any normal person would have take 3-4 hours, minimum - futzing with google sheets and such (lots of task/taskgroup independent counts). Even I would have taken a minimum of 10 minutes. Cowork had it for me in under 30 seconds - I didn't even have to really look away from the zoom.

I haven't use the Atlassian Jira interface to manage my sprints/issues in 6+ weeks. All that cognitive drain is gone - I just ask CoWork to update the sprints, link issues to change-notices, comment and change status, set epics, etc... Zero need to find the damn field in Jira.

CoWork is legit.

So - presumably you have another site in mind that is better. I'd be intrigued to know which one you would recommend.

I'm guessing my brain has been permanently warped by working in a hashistack environment for the last 5+ years. All of these KV pairs are always either in a consul (bog-standard config information) or vault (secrets/credentials) in a consul-template associated with the job - and rendered at run time. And, of course - critically - the consul-template updates those credentials and values as they are routinely rotated.

I did a ^F for "update" and "dynamic" - didn't find any hits in that doc. Managing your dynamic environment values is table-stakes in most large deployments - and I didn't see any reference in `secretspec` as to how they plan to do that.