ademup

Faster inference would solve nearly all of the issues I have with AI models. I really liked working with Opus 4.8 and I would greatly prefer a 100x faster version of it, with 100x the tokens, at the same price, to any of the openai or anthropic models that have come out since.

I really love this idea given the recent controversy around mathematics solutions. It seems like a "mere suggestion" of success has a positive impact on finding solutions. We know this technique works in humans, from which this is all derived from, so it seems to make sense.

Great news, then! TFA: "Last week, we previewed the Model Hardware Standard, which allows Claude to directly and safely operate laboratory equipment."

I'll probably be downvoted to the bottom of the ocean, but I feel this needs to be said:

- Why does it matter who or what wrote a thing? And how would you know if the /content/ is worthwhile, unless you read it...

- The ability to "detect AI" is imperfect at best. 90% AI written? 5%? How would you know, unless you read it....

- If you didn't read it, then why brag about it with an "AI;DR"?

- If you are wrong, and you announce it to the world, how would it feel to someone who spent a lot of time and energy writing?

A single counter example: In East Bay CA. I designed and built my 9kw solar roof installation myself. No electricians at all. The process was very easy and working with my city and power company was simple, with no blockers at all. 4 years running. On a yearly basis, we make over double what we use. It boggles my mind how solar isn't on every house in the country.

Yes apparently so: my customer is a very intelligent research director at a well known company. He was very excited to tell me about going "headless", and I was very confused for a solid 5 minutes before figuring it out. He told me that this was the term his vendors were using.

My clients and colleagues are bragging about "headless web sites", by which they mean "sites built without WordPress (and kin)", by which they mean "sites built entirely using the English language via AI"

In that sense, the author might be implying that "web framework" may soon be a thing relegated to history

Downvoted for breaking this guideline: "Don't be curmudgeonly. Thoughtful criticism is fine, but please don't be rigidly or generically negative."

The "It's clickbait" comment at the end made me feel pain for the site buider, and I didn't even put any work into the website. They made a thing and put it out in the world. Some people like it: as evidenced by other comments here.

That they mention the top of the price range instead of the bottom lends a lot of credibility to my mind.

Agreed within the narrow confines of web dev (which is all that I do). I used to write 500-1000? LOC per day but now I've built several full fledged (250k+ arr) sites with more features than I've ever been able to implement in such a short time: all without editing a single line of code.

My guess is they are still very useful for more difficult code! But yeah, I can't imagine ever caring about "code" any more, and therefore cannot fathom the need for a full fledged IDE.

Thank you for this post! 1) I've been considering Zed for a long time, but it hasn't worked well in my KVM. Due to the poor(?) ability to pass GPU through on my Ubuntu 24.04 machine. I have read that 26.04 may have fixed this so I'll try it again!

2) I am in the same boat with slowness. I've been using PHPStorm for over 10 years and it has always been "slow", but the newest pain point is that I will have claude in a terminal update a file. If the file is open in PHPstorm's viewing pain, it might take 10+ seconds to update the contents: I now always "update from disk" if I want to copy the contents outside of my KVM. It's just absolutely terrible workflow.

3. I have also found all of their AI efforts to not only be poorly executed, but executed in poor taste: it's just IN THE WAY rather than being helpful.

4. I mostly don't appreciate most of their features, generally. My flow is pretty simple. I no longer use most of the features. I just don't need the 8000lb elephant any more.

This on a 64GB ram Ryzen 7 5825U.

I call this the "Everything in Moderation" fallacy. From what I've heard people who say it, they emphasize the everything part of it. In other words almost everything is bad for you so just eat a little bit of everything and you won't get too much of the bad stuff. It's maddening.

Respectfully disagree. An AI with full access to robots could do everything on its own that it would need to "survive" and grow. I argue that humans are actually in the way of that.

For sure! But just for reference, I'm on a mid-tier 2022 Dell Inspiron laptop: Ryzen 7 5825U with 64GB ram and 500GB SSD.

On it, I run Ubuntu 24.04 as my host, and my guest is Lubuntu with 16GB ram and 80GB ssd for my KVM.

I almost always have 2 instances of PHPstorm open in both Host and and Guest with multiple terminal tabs running various agentic tasks.

I'm surprised to see so many people using containers when setting up a KVM is so easy, gives the most robust environment possible, and to my knowledge much has better isolation. A vanilla build of Linux plus your IDE of choice and you're off to the races.

For anyone else on the fence about moving to CLI: I'm really glad I did.

I am converting a WordPress site to a much leaner custom one, including the functionality of all plugins and migrating all the data. I've put in about 20 hours or so and I'd be shocked if I have another 20 hours to go. What I have so far looks and operates better than the original (according to the owner). It's much faster and has more features.

The original site took more than 10 people to build, and many months to get up and running. I will have it up single-handedly inside of 1 month, and it will have much faster load times and many more features. The site makes enough money to fully support 2 families in the USA very well.

My Stack: Old school LAMP. PHPstorm locally. No frameworks. Vanilla JS.

Original process: webchat based since sonnet 3.5 came out, but I used Gemini a lot after 2.5 pro came out, but primarily sonnet.

- Use Claude projects for "features". Give it only the files strictly required to do the specific thing I'm working on. - Have it read the files closely, "think hard" and make a plan - Then write the code - MINOR iteration if needed. Sometimes bounce it off of Gemini first. - the trick was to "know when to stop" using the LLM and just get to coding. - copy code into PHPStorm and edit/commit as needed - repeat for every feature. (refresh the claude project each time).

Evolution: Finally take the CLI plunge: Claude Code - Spin up a KVM: I'm not taking any chances. - Run PHPStorm + CC in the KVM as a "contract developer" - the "KVM developer" cannot push to main - set up claude.md carefully - carefully prompt it with structure, bounds, and instructions

- run into lots of quirks with lots of little "fixes" -- too verbose -- does not respect "my coding style" -- poor adherence to claude.md instructions when over half way through context, etc - Start looking into subagents. It feels like it's not really working? - Instead: I break my site into logical "features" -- Terminal Tab 1: "You may only work in X folder" -- Terminal Tab 2: "You may only work in Y folder". -- THIS WORKS WELL. I am finally in a "HOLY MOLLY, I am now unquestionably more productive territory!"

Codex model comes out - I open another tab and try it - I use it until I hit the "You've reached your limit. Wait 3 hour" message. - I go back to Claude (Man is this SLOW! and Verbose!). Minor irritation. - Go back to Codex until I hit my weekly limit - Go back to Claude again. "Oh wow, Codex works SO MUCH BETTER for me". - I actually haven't fussed with the AGENTS.md, nor do I give it a bunch of extra hand-holding. It just works really well by itself. - Buy the OpenAI PRO plan and haven't looked back.

I haven't "coded" much since switching to Codex and couldn't be happier. I just say "Do this" and it does it. Then I say "Change this" and it does it. On the rare occasions it takes a wrong turn, I simply add a coding comment like "Create a new method that does X and use that instead" and we're right back on track.

We are 100% at a point where people can just "Tell the computer what you want in a web page, and it will work".

And I am SOOOO Excited to see what's next.

The status quo system you describe isn't objectively optimal. It sounds archaic to me. "We" would never intentionally design it this way if we had a fresh start. I believe it is this way due to a meriad of reasons, mostly stemming from the frailty and avarice of people.

I'd argue the opposite of your stance: we've never had a chance at a fresh start without destruction, but agents (or their near-future offspring) can hold our entire systems "in nemory", and therefore might be our only chance at a redo without literally killing ourselves to get there.

Is Ubuntu 24.04 supported? (Docker Desktop doesn't support 24.04 currently)

Bulma.io layout is broken in landscape mode on my phone (responsively and full page refresh) with several sections going beyond the edge of the page. Firefix on Pixel 7.

I'd love to use this because it looks good, but this makes me lose trust.

For a far more in depth look into high intensity gardening, with real data and realistic expectations, I strongly recommend John Jeavons: How to Grow More Vegetables: Than You Ever Thought Possible on Less Land Than You Can Imagine

Location: San Francisco Bay Area

Remote: Remote, Hybrid, On Site

Willing to relocate: No

Technologies: Linux, PHP, MariaDB, VPS, Javascript, HIPAA, CMS/EHR creation, APIs, Analytics, Bootstrap, Plotly, Twilio, Sendgrid. Small amount of Python.

Résumé/CV: https://www.linkedin.com/in/ademmiller/

Email: adem@mentalnexus.com

- Over 15 years of experience writing multiple CRM/EHR web apps in PHP and JavaScript from the ground up without frameworks.

- 8 Years as Co-Founder and CTO orchestrating millions of surveys in the private pay mental health and addiction treatment arenas

- Extensive experience working with billion dollar valuation companies as customers and integrating with the wide range of software vendors and APIs they consume

- Degrees in Business Administration and Small Signal Electronics - Huge fan of entrepreneurs and small startups

I'm experiencing the same massive pain point from Twilio. Our business sends survey links to about 300 people a day who have all opted in. We've been doing this for 7 years. Our undelivered rate had been under 5% for years. It's now at 33% and is threatening our business.

Twilio's support has been absolutely dismal. They say things like "Your number is not attached to your campaign" and show a screen shot from their admin screen, but when I point out it is all connected in the console: they respond with the generic Twilio links about A2P implementation.

Has anyone figured out a solution? ... I'm happy to switch to another company at this point.

Can you please point out where the post judges people for using screens? I see "don't use screens here", which isn't a judgment of people who use screens, it's just a rule. Your statement, however, is very judgemental. Especially of a place you have never been and appear to have decided to never visit. For clarity, here's what I consider the harshest judgemental part you posted: "Can you imagine the conversations you'd find yourself in at this place, given this is how the "owner" sees the world?"

Finally, I'm pretty sure the original article is some sort of creative prose or story, given the rest of the cmart site. If so, then its appears to have illicited strong reactions, which may be a sort of success in itself.

I had low diopter astigmatism and got lazik 6 years ago at 39. I consider it the best decision I've ever made that was just for me. I still have perfect vision, and if it degrades again I will happily get it done again. Just my experience, but I have yet to meet anyone who has regretted it.

I really like the concept and experience even though I almost missed out! I thought it was _only_ available as an app when I loaded the page on my phone. I rarely (basically never) install apps so I closed it immediately. It wasn't until I read on HN that I realized there's a web version.

Your story sounds similar to mine, and your choice to use HTMX has me motivated to check it out. The sum total of my software supports 5 families' lifestyles entirely on LAMP MPAs with no frameworks at all. Thanks for posting.

The article starts off with what I believe is a faulty premise, that churn "is" rejection. Trying something new is such a deeply ingrained part of our DNA that there willl always be a segment of your users and staff who do so regardless of all of your attempts to prevent it. It isn't rejection so much as standard human behavior. Also, I thing "rejection" is far more inflammatory than "churn".

I'm not sure why others downvoted this comment, but I did so because of the bit about "resistance to censorship ... Hash power" is so incredulous to me as to approach a dangerous level in its appeal to the vulnerable. Ie it reeks of proselytization.

I would really like to know why this comment was down voted. It is a perfectly valid question with a rational explanation. Indeed, I was surprised to have needed to scroll down so far to find it, as it was the very first question that popped into my head as well.

Down voted because the entirely of this message is directed at the poster and nowhere near the topic.

For me, coffee tastes far superior when it is much more diluted than "normal". 1 liter French press with 2 tablespoons of grounds. Then put about 2 tablespoons of this liquid with 32 oz of ice water. It should now look like light tea. It's a refreshing beverage I can drink from 8am to 2pm with no downsides. This drink can be replicated at starbucks as "can I get a shot ... And a venti ice water with light ice". Then combine. Side benefit is they often charge much less than an Americano.

My batches of french press last 2 days, so in this way I suppose I am reusing the grounds.