| Apr 03 |
1. Go to openrouter
2. Get a free account and free api key
3. Check their Claude code docs & update your env variables accordingly
4. Set the model to qwen 3.6
5. Enjoy Claude code with an opus level model with 1m context window for free
✅ Problem solved
|
2 |
0 |
0 |
631 |
259 |
. |
| Apr 01 |
I run a software agency and @claudeai just cancelled our entire team with no warning and no reason why. @trq212 @bcherny what can I do here? Can you even tell me what the reason is so we can fix? This is massively disruptive to business.
There goes my day 🙃 https://t.co/oYEePLz9gy
|
2 |
0 |
3 |
874 |
283 |
. |
| Mar 31 |
What a shit show - hundreds of @claudeai code users getting their account suspended with no recourse https://t.co/7dcILK3kd1
|
3 |
0 |
0 |
621 |
124 |
. |
| Mar 31 |
Have seen this happen a few people in the feed already, now it's my turn. No explanation, no warning, using claude code normally. Why make people reliant in that way then completely destroy trust. Ugh
@trq212 https://t.co/OtXm6xH6Hu
|
0 |
0 |
0 |
903 |
233 |
. |
| Mar 08 |
This 100%. @Replit used to be my favourite product philosophically - I think I built one of the first rails-compatible boilerplates for it & was tweeting about it back in 2020.
But every time I hear Amjad speak nowadays it just feels so inauthentic. I know it's tricky to find PMF & you have to sell the dream etc. but I don't think I've heard him be humble or radically honest on any pods - it's all stories of how he had this vision & did crazy hard things & hosts blowing smoke up his ass when in reality they've pivoted countless times & still don't really have a clear vision for where things will end up (which is fine, just be honest about it).
|
4 |
0 |
0 |
626 |
652 |
. |
| Feb 05 |
This is so impressive, @kamrify is the gold standard of solo product builder - haven't seen anyone else who is as good at everything - design, ux, engineering, even the product walkthrough is great.
|
4 |
0 |
1 |
645 |
198 |
. |
| Feb 04 |
So... using the super fast inference providers - Groq & Cerebras - is kind of addictive.
So addictive that even though they make silly mistakes & require fixing, I think I prefer having to do multiple passes than to wait 2 to 3 minutes for Opus on Claude Code.
|
1 |
0 |
0 |
815 |
270 |
. |
| Jan 27 |
Been hooking up a lot of tools to Claude Code this week, and my conclusion is that just having a credentials folder and telling it to create a skill for each new integration, with the API endpoints, is more reliable than the official MCP, still not seeing the value in MCP
|
7 |
0 |
1 |
689 |
272 |
. |
| Jan 17 |
Raw model intelligence still has close-to-zero common sense. If you give it a harness and the right context, it's incredibly powerful - the big question is will those harnesses be able to add the "G" to AGI outside specific vertical domains
|
4 |
0 |
3 |
823 |
240 |
. |
| Dec 06 |
This is true if…
- You are building very highly interactive frontends
- You don’t want to spend any time figuring out a playbook/set of patterns (which is completely fair and true of 99% of people)
It’s not because of any fundamental difficulty with doing this stuff without js, It’s just that js libs have consumed everything so there’s not as much material out there.
Because of this I don’t agree with the “you’ve got to be a masochist” framing. You just want to accept more up front work for lower long term complexity
|
5 |
0 |
1 |
923 |
525 |
. |
| Nov 16 |
There will be movies made about this werk and this man. You couldn’t script it, incredible scenes
|
3 |
0 |
0 |
793 |
97 |
. |
| Nov 09 |
This is the sweet spot for me when building components with more-than-basic interactivity: A Minijs component that uses preact standalone for templating.
- Easy to follow: Very little framework-specific knowledge needed.
- I can do simple slot-based composition in the dom. e.g reorder/remove parts.
- I can use jsx for templating (technically htm but looks and feels the same.
- Super performant
- No build step
|
3 |
1 |
0 |
720 |
416 |
. |
| Nov 02 |
This is still the best quick-website-builder I've used - really really good and not as well-known as it should be. You should check it out!
|
7 |
2 |
1 |
882 |
139 |
. |
| Nov 01 |
For people building software like this, where does design fit in? There’s no way the LLMs are giving smooth ui/ux on the first pass, is there?
|
1 |
0 |
3 |
534 |
142 |
. |
| Oct 15 |
Couldn't make this up - logged on this morning excited to try out some new AI data analyst tools - *both* of the ones I've heard of have critical bugs on the onboarding flow https://t.co/7bdtlXKXf3
|
1 |
0 |
0 |
897 |
197 |
. |
| Oct 13 |
I love that the remix guys are popularising web standards again among js folks, but I (currently) feel that the improvements they’ve made are in a dead zone. Marginally better than react, but not a step change in simplicity or functionality yet
|
2 |
0 |
0 |
849 |
244 |
. |
| Oct 13 |
Can someone tell me who all these tweets are subtweeting?
|
0 |
0 |
1 |
698 |
57 |
. |
| Oct 10 |
Another paper on HN today highlighting the shortcomings of Reasoning Models.
My synthesis (Ironically summarised with Claude):
- Humans use deductive reasoning with state tracking - when you rule something out, it stays ruled out in your working memory. Your mental model updates and enforces consistency - ruled-out causes become hard constraints on all future reasoning.
- LLMs use sequential probabilistic reasoning - they can rule something out but may "forget" and re-consider it later because they lack hard constraints.
- LLMs have soft probabilistic weighting instead of explicit working memory - earlier conclusions influence but don't guarantee exclusion.
- Both are "reasoning", just with different architectures and failure modes.
Can LLMs ever do perfect logical reasoning?
- Pure LLM architectures will likely always have limitations because they generate text token-by-token based on probabilities, not by executing logical operations
- They lack explicit data structures for tracking ruled-out hypotheses or maintaining symbolic state between tokens
|
0 |
0 |
1 |
665 |
1.1k |
. |
| Sep 26 |
Every single time there's controversy on the timeline, I somehow only feel compelled to share the viewpoints of the side I mostly *don't* agree with, because the side I agree with is so over represented, finding a good point from the other side feels like it should be amplified
|
2 |
0 |
0 |
650 |
278 |
. |
| Sep 22 |
For years I listened to the twitter sentiment of "Don't do consumer it's just too hard". Then me & gf decided to give it a shot. Since then we've done 40m views on Tiktok and are averaging 30 new customers/day.
Don't believe all the advice you hear on here. https://t.co/46GKmnelpe
|
9 |
0 |
1 |
506 |
259 |
. |
| Sep 21 |
1. Get a VM from contabo or hetzner
2. Connect from editor via remote-ssh
3. gem ‘passenger’ & bundle install
4. passenger start —daemonize
5. Add dns record for VM’s ip address
6. One line in Caddyfile
7. ‘service caddy restart’ provisions SSL
You now have a “live” rails app
|
1 |
0 |
0 |
979 |
283 |
. |
| Sep 18 |
.@RoamResearch is by far the nerdiest product I use and I love it
|
4 |
1 |
0 |
786 |
65 |
. |
| Sep 15 |
Don't meet your (twitter) heroes folks
|
2 |
0 |
1 |
562 |
38 |
. |
| Sep 15 |
One week ago: MCP is literally just over complicating a rest API
Two days ago: Oh wait, maybe I'm wrong, look at all these cool connectors. Maybe they can do cool stuff
Today: MCP is literally just over complicating a rest API... https://t.co/3mG7RYycjW
|
7 |
1 |
0 |
592 |
231 |
. |
| Sep 13 |
https://t.co/jM7vZ3Tq7K
|
6 |
0 |
1 |
930 |
23 |
. |
| Sep 11 |
AI Products that have *stuck* for me so far
- @claudeai as daily chat assistant
- @WisprFlow for dictation
- @codegen for code review
- @magicpatterns for interactive prototypes
I also use Cursor, but mainly for auto-complete and deep-diving, and it doesn't feel like it has the same delight and consistency as the other four
|
4 |
1 |
0 |
521 |
328 |
. |
| Sep 10 |
Nikita’s been good so far but pity he’s dismissing this. If you want a platform everyone posts one you need to give them an incentive. Lack of engagement is a massive issue - only reason I’m still here is because I’ve already built a (tiny) following, no way I’d stick with it if I started today
|
1 |
0 |
2 |
568 |
295 |
. |
| Sep 04 |
Got to Rails World, the main area was so quiet, turns out everyone was at the Omarchy talk, which is overflowing https://t.co/Kb4iTkiV7P
|
10 |
1 |
1 |
635 |
136 |
. |
| Sep 01 |
Who's coming to Amsterdam this week for @railsworld?
|
13 |
0 |
5 |
903 |
52 |
. |
| Aug 24 |
Re: App Mafia hate, as someone who’s skeptical of info products. Myself/gf listened to pod with the same guys, took a shit ton of notes. 10 wks ago launched an app with their exact playbook - now doing avg 1k+/day. To be fair they do know what they’re talking about & it does work
|
5 |
0 |
0 |
526 |
284 |
. |
| Aug 21 |
Can someone make this but in a way that’s socially acceptable for adults plz. I’d be fit as fuck if my gym had fun stuff like this. https://t.co/41Stvnyehr
|
5 |
0 |
0 |
505 |
155 |
. |
| Aug 19 |
I think I’m gonna have to go build an html first vibe coding platform to avoid people being sucked in by takes like this
|
12 |
0 |
3 |
665 |
120 |
. |
| Aug 13 |
Love it when plans start to come together https://t.co/VDCmUnhYsE
|
5 |
0 |
2 |
915 |
65 |
. |
| Aug 13 |
The more I've thought about it, the more I think the obsession with avoiding *any* excess kilobytes on first page load is silly for most web apps (excluding maybe e-commerce).
Web apps are designed for long running user sessions: A slightly longer-than-usual first request is 1) Not noticed by 99% of users, and 2) A reasonable way to give your browser the code it needs to run the rest of the web app.
(Talking about a few hundred kbs of css/js here, not MBs). On a half decent connection with server round-trip you're looking at a few hundred milliseconds. On very slow connections it's a few seconds, which is bad, but in that case the app is going to need all sorts of other optimizations, and even then, front-loading the assets often still comes out as best all-in solution, assuming you're not using a ton of libraries, which you shouldn't be in most cases.
|
3 |
0 |
3 |
628 |
867 |
. |
| Aug 09 |
https://t.co/aSBC2Ao3k6 https://t.co/fpIPFXYyzs
|
1 |
0 |
0 |
549 |
47 |
. |
| Aug 09 |
I really need to share my side projects more.
I've seen a few people ask for a no-build version of Tailwind recently - I happened to build one a few years ago that still works pretty well - link below
|
4 |
0 |
2 |
959 |
202 |
. |
| Jun 27 |
Interest piqued with the first line - “oh wow someone managed to vibe code an app that makes real money”
Nope, turns out its a “$1.7m Saas” because that’s what replit told him it would cost to build it 🙃
|
10 |
0 |
1 |
617 |
204 |
. |
| Jun 16 |
I wish there were some kind of universal way to measure software developer output for hiring - even if somewhat imperfect
Interviewing a lot recently and seeing candidates that look identical on paper but one of them can complete in one day, what takes the other 2 weeks.
|
0 |
0 |
1 |
619 |
272 |
. |
| Jun 12 |
Would love to know how many Twitter users have been “Guillermo’d”
When the legend slides in your DMs and makes you immediately take your shit more seriously
|
6 |
0 |
1 |
661 |
158 |
. |
| Jun 05 |
When you transfer $750 in 3 seconds for a fraction of a cent, it's hard *not* to see the 3% that payment companies extract as anything else but extreme greed. https://t.co/dLl0RhOck5
|
3 |
0 |
1 |
569 |
158 |
. |
| May 28 |
Just had a long conversation with Claude trying to guess what kind of patterns @remix_run is working towards based on their blog post.
Extreeeemely intrigued https://t.co/YrnSD76je5
|
5 |
0 |
1 |
581 |
183 |
. |
| May 26 |
The most pessimistic take on AI assisted software development is...
There's a fundamental ceiling on how well apps can be built if there is no feedback loop (the coder can see and interact with the results of the code).
A feedback loop requires reading and understanding the browser (and/or native apps).
This presents 2 problems
- The state of the art in browser interpretation even today, is still very low reliability.
- Most apps have an obfuscation layer (react/next etc) between what's in the codebase and what the browser renders, making this harder.
There is a real possibility that things improve, but don't reach to the point that even mid level developers can be removed.
|
0 |
0 |
0 |
671 |
687 |
. |
| May 08 |
😂 some people can't help themselves once they catch the faintest scent of Schadenfreude. The 37signals owners have more wealth (measured in dollars and freedom) than 90% of founders, which is exactly why they indulge in projects like this. They’ll be just fine 😂
|
7 |
0 |
1 |
849 |
262 |
. |
| Apr 29 |
This is exceptionally dumb. @codeiumdev @cursor_ai @cline will you guys figure this out? First mover gets to define the standard a la MCP https://t.co/yGgYbyw3Gj
|
6 |
0 |
7 |
673 |
161 |
. |
| Mar 29 |
Just realised. I've been working on a few UI components for several hours - copy-pasting the entire file into Claude repeatedly because it's faster than using cursor, and there hasn't been a single instance where it's returned broken code or added unwanted code out of 50+ passes
|
4 |
0 |
2 |
905 |
279 |
. |
| Mar 26 |
Biggest update to my productivity recently is @foragemail - simple idea, well executed
- Sign up with gmail
- Now all transactional emails skip your inbox
- One email per day with a digest of everything
I'm actually responding to emails again. Highly recommended.
|
3 |
1 |
1 |
583 |
265 |
. |
| Mar 21 |
It feels like the quality bar for new digital products is in decline. I like to try new stuff, I think at this point I'm averaging one product a week where I have to reach out to support to just get the first run to work. Mostly AI first products which are harder, but still.
|
3 |
0 |
2 |
908 |
275 |
. |
| Mar 11 |
For what Dario’s saying to be accurate, one of the following would have to be true.
- Dario is exceptionally disconnected from reality
- Dario is very dishonest and is saying this intentionally
- Anthropic are about to drop the biggest leap in coding capabilities since GPT 3
|
8 |
0 |
1 |
583 |
277 |
. |
| Mar 11 |
Announcing Scout: AI that can do more than just chat.
A few months ago a major milestone passed (somewhat) quietly: AIs can now competently control a web browser & reliably complete multi step tasks.
Now that the tech is there, there are a few products in the space, but we're still in the extremely early stages of figuring out what's possible and building the right application layer.
The best products will come from 1. Great product taste, 2. The ability to combine and deploy the existing technologies (models, libraries), and 3. The ability to ship-iterate-improve to find the right patterns & UI paradigms.
We've built a team at @rlygoodsoftware that has all of those characteristics, so rather than wait for someone else to build the product we want, we've begun working on something I'm super excited about.
Some things we're focusing on:
- Task Focussed (as opposed to chat-centric): With Scout you can quickly fire a task off or run it on a schedule - you don't need to monitor and respond to a chat (but Scout will ping you if it needs you).
- Credential Management as a first class citizen: No pasting your password in plain text into a chat. Scout has a built-in password manager that lets you save and pass your credentials safely to the Agent. We also have plans to let you deploy on your own cloud.
- Fine grained Control: You can set Scout to be Cautious, Careful, or Confident. In cautious mode, Scout won't proceed without asking for signoff on any PQAs (Potentially Consequential Actions) like making payments or sending messages. Confident mode - if you let it - can buy your groceries and have them delivered without your input.
These are just some of the things we're exploring. We've already built a V1 and are testing internally for now, but would like to get more people using the product from April. I'll also be sharing some sneak peeks on here as we build it out
If you're interested head over and sign up to the waitlist - link in next tweet.
|
14 |
0 |
1 |
692 |
2k |
. |
| Mar 08 |
Have tried all the landing page builders and @pagyco is still the best for quickly throwing something together. So easy to use, great work @hernansartorio 👏
|
8 |
3 |
1 |
686 |
156 |
. |
| Feb 24 |
As someone who has opened entities in Asia, Europe and the US, generally speaking the logistical price-of-entry to run a business is very high, regardless of country.
People pointing to marginal difference in ease-of-opening-a-business are missing the cost of *ongoing* requirements in accountants, bookkeepers, compliance.
Yes EU is bad for startups and US generally has fewer regulations.
But for the average business, the tax-on-entrepreneurship of not having a zero-cost way, provided by the government, to stay compliant, is what we should be complaining about, in *every* country.
|
2 |
1 |
1 |
734 |
592 |
. |
| Jan 20 |
Looks like no obvious solution in 2025 that does "Sign up, Add Funds, Send SMS reliably, internationally with a single API call."
All google-page-one products make you treat each country separately - separate pricing, often have to buy one number for each country.
🤷♂️
|
0 |
0 |
1 |
773 |
271 |
. |
| Jan 19 |
Experience with @VonageDev:
- Add a card, add credit, confirm payment
- Vonage rejects the payment
- Try another card (total of $20)
- "You have reached your daily Vonage API spending limit" - balance is still 0
If you want to be known as "easy to get started with", this is the most basic stuff.
Any other good, simple SMS providers out there?
|
1 |
0 |
1 |
604 |
349 |
. |
| Jan 18 |
Whenever I read things like this it reminds me how little CEOs at big companies understand what’s happening in their own company and how unreliable anecdotes are when they go beyond one degree of separation
|
2 |
0 |
0 |
562 |
206 |
. |
| Nov 10 |
If you want to build a payments product like this today, your options are:
1. Pay a company like Stripe a percentage fee (anywhere between €100 to €200 *per transaction* for a €10k payment.
That's it - there are no more options
Online payments are not anywhere near solved https://t.co/sVpT7IpASt
|
2 |
0 |
1 |
500 |
299 |
. |
| Nov 03 |
90% of the time I see the term “Skill issue” these days, it’s used in bad faith.
|
4 |
0 |
4 |
603 |
80 |
. |
| Oct 29 |
Highest leverage use of Claude/Cursor so far:
Spent ages trying to find a good web component version of choices.js/select2/tomselect - couldn't find anything
v0 in Claude Artifacts took 15 mins, ran out of context so switched to Cursor. 45 mins later & have prototype + docs https://t.co/FLVfdJrxia
|
1 |
0 |
0 |
632 |
305 |
. |
| Oct 23 |
Another data point here:
CSS went through a build-time library phase (scss, sass etc), where it was the norm to use a higher level language and a build step because CSS couldn't do things like variables, nesting, calculations
Now CSS has those things, and we're removing these build time libraries wholesale.
Won't this also be true for js meta-frameworks at some point? If so, when?
|
5 |
0 |
0 |
729 |
387 |
. |
| Oct 19 |
K fuck it, gonna also build this in html first rails as a comparison.
|
8 |
0 |
1 |
542 |
69 |
. |
| Oct 18 |
What are the arguments that declarative frameworks (React, vue, svelte etc) buck the "Obsoleted by The Platform" trend?
|
1 |
0 |
1 |
676 |
119 |
. |
| Oct 17 |
You'll never convince me that this is how we will be writing/extending html in 5 to 10 years time. https://t.co/usiS76zTSM
|
9 |
0 |
3 |
849 |
122 |
. |
| Oct 14 |
This paragraph is something else. https://t.co/7QWl4FrZSx
|
9 |
0 |
2 |
868 |
57 |
. |
| Oct 11 |
Put me in the *extremely skeptical* group for this happening. Many other ways AI will disrupt things, but humans mostly want their tools to be predictable, so I don’t think we’ll see shapeshifting software become widespread.
|
0 |
0 |
0 |
527 |
225 |
. |
| Oct 06 |
Doing more cursor-coding recently - an observation:
It's difficult to switch brain-modes between tell-cursor-what-to-write and write-the-code-yourself, even if the latter is actually quicker than typing out a prompt.
Assuming I'm not an outlier, I *think* this will mean people who go the ai-assisted route will tend toward writing less and less code over time, eventually approaching zero.
Not sure how I feel about that yet.
|
7 |
0 |
3 |
865 |
431 |
. |
| Oct 06 |
Love this concept. I’m a leverage programmer for sure.
|
4 |
0 |
0 |
891 |
54 |
. |
| Oct 05 |
I’m in SF and Waymos are EVERYWHERE and every time I see one I tilt my head to check if there’s a driver and the wonder and amazement of there not being one still hasn’t worn off
|
9 |
0 |
2 |
586 |
178 |
. |
| Oct 03 |
For example to build one of the form flows @shl mentioned, the solution didn't really require a lot of custom htmx. It was, roughly speaking:
- Add hx-boost to the body and use plain old Rails form helpers.
- Use a css snippet and hx-indicator to show loading spinners on link clicks and form submits.
- Use hx-preserve to prevent elements being replaced on new page load.
But it takes quite a long time to figure those things out.
|
10 |
1 |
2 |
940 |
435 |
. |
| Sep 30 |
Unless something has changed in the last year Heroku is still the all round best provider for medium-traffic apps even at those prices. Provided you swap out heroku postgres for RDS and limit addons
|
4 |
0 |
1 |
657 |
198 |
. |
| Sep 27 |
Have had multiple interactions in the last few days (on here and at Rails World) that make me think maybe I should do a little htmx-on-rails course - there’d also be heavy overlap with Turbo because it’s ultimately about “thinking in hypertext” as it applies to rails
|
6 |
0 |
1 |
896 |
267 |
. |
| Sep 23 |
Watched Caleb's flux demo and had a few thoughts.
1. Wow, we have basically the exact same primitives for our in-house css system (minus the layouts).
2. 95% of the value is in the CSS - coupling it to Laravel/Livewire seems limiting when it could be sold as CSS only and slot into anyone's stack.
3. Maybe I should sell ours too 🤔
|
4 |
0 |
1 |
530 |
332 |
. |
| Sep 05 |
You either die a tailwind user, or live long enough to end up building your own css ui kit 🙃
|
2 |
0 |
0 |
700 |
92 |
. |
| Aug 27 |
My impression is Rails & Laravel are the most productive full stack, solo-dev frameworks out there, and that Laravel is much farther ahead from an ecosystem/batteries included perspective.
Ruby is just a much nicer language (in my subjective opinion) than php
|
3 |
0 |
1 |
578 |
264 |
. |
| Aug 13 |
In what world is this not the absolute pinnacle of rent seeking? How is this considered normal in 2024?
|
1 |
0 |
2 |
504 |
103 |
. |
| Aug 06 |
My issue with payments companies (including @stripe) is this exactly. The thinking seems to be “we wired together a dysfunctional system, therefore we deserve a percentage of all transactions” even though today there are systems to transfer value much much cheaper, and one of the things preventing their adoption is the preservation of the old system by the payments companies who built an empire around them.
|
2 |
1 |
0 |
553 |
412 |
. |
| Jul 31 |
Hot Take: Hacker News is right on this one.
|
3 |
0 |
2 |
573 |
43 |
. |
| Jul 16 |
I work fairly long hours and haven't had a proper holiday in years, but I can decide to go to the cinema with the family at 3pm on a Tuesday or hit the road on a whim, and it's an amazing feeling coming from 10+ years of startup life.
Being a business owner is tough for the first few years but if you stick it out the freedom is unmatched.
|
6 |
0 |
0 |
746 |
341 |
. |
| Jul 14 |
I generally try to be nice online, but I *despise* these kinds of “You’re not smart enough to sit with us *real* engineers” tweets
|
16 |
0 |
2 |
928 |
130 |
. |
| May 30 |
I think @t3dotgg is mostly bad faith, but there is a pretty important point here. The html-over-js-brigade (of which I'm a part) should have better responses and resources to point to when jankiness is raised.
In my experience It's 100% possible to make server-rendered apps that *feel* as good/fast as SPAs. But it does require much more intentionality than is involved if you take a declarative js library off the shelf.
|
1 |
0 |
1 |
913 |
423 |
. |
| May 29 |
A few weeks ago @RevolutApp started bouncing every payment to our Filipino team, who we’d been paying for months, and blamed it on “an issue on the other side” with no other explanation and no recourse. How can there be so little accountability for something so important 🤯
|
1 |
0 |
2 |
700 |
273 |
. |
| May 17 |
The secret to getting a payment provider to provide you with basic communication/support: Have your complaint post get nearly a million impressions. Easy! https://t.co/EDolCHbZCZ
|
2 |
0 |
1 |
584 |
178 |
. |
| May 07 |
Just found out that yc is back to in-person since '22, and now I feel bad for the lukewarm advice I gave to people on whether they should apply (I was in the remote batch and was quite disappointed). If I said this to you you can safely ignore - I'm guessing an in-person yc would https://t.co/uj7WTjQtWo
|
6 |
0 |
1 |
681 |
298 |
. |
| Apr 29 |
Some great single-tweets in this thread
https://t.co/4HEFL1fg8I
|
2 |
0 |
3 |
506 |
64 |
. |
| Apr 24 |
It's interesting how people can interpret the same thing so differently.
My experiences - personal, family, friends - is that it's basically impossible to get half-decent, reasonably affordable therapy - not the thing I'd be calling for *more* regulation on.
Personally don't think the AI chat tool is appealing. If it were the case that the alternatives were abundant and affordable then *maybe* would say this kind of thing is harmful. Otherwise this seems like a good example of "It's the worst thing we have, apart from every alternative"
|
0 |
0 |
2 |
923 |
546 |
. |
| Apr 23 |
Glad I didn’t see these types of tweets when I was thinking of applying to yc. The one day of prep for runway for 9 months when I was pre revenue at a very generous valuation was one of the highest leverage things I ever did. Most people’s early stage ideas are worthless.
|
2 |
0 |
0 |
540 |
272 |
. |
| Apr 18 |
This is going to be like crack to people who have lost loved ones and want to talk to them again
|
6 |
0 |
2 |
842 |
96 |
. |
| Apr 15 |
Writing something this polished in this amount of time/code is *so* satisfying. No build step, and full control of the markup. 🤩 https://t.co/qdzgDKWMGf
|
5 |
0 |
1 |
613 |
152 |
. |
| Apr 10 |
The optimum amount of conflict in friend/family relationships is not zero - it's being able to get a little cranky at each other and have it not boil over. Seen so many examples now where people have zero conflict for years then something happens and they never talk again
|
8 |
0 |
2 |
833 |
272 |
. |
| Mar 27 |
Shopify's AI helpbot is not very helpful🤦♂️
Related: Every AI driven chatbot interaction I've had so far is similar to this https://t.co/oXY0UF8wY5
|
1 |
0 |
1 |
555 |
149 |
. |
| Mar 25 |
Is there a ruby/rails version of https://t.co/OUtr9ZEFPP?
Love the interactivity & gamification, think it could be incredible for getting juniors up to speed.
|
1 |
0 |
1 |
797 |
162 |
. |
| Mar 19 |
Latest attempt at explaining the core Mini js concepts in the simplest way possible. https://t.co/QXC4CjAXu9
|
3 |
0 |
2 |
980 |
108 |
. |
| Mar 19 |
This is a fascinating exploration of an alternative way to “write code”
|
3 |
0 |
1 |
678 |
71 |
. |
| Mar 16 |
Who else is making music (or anything) that gets this kind of reaction, not just from die hard fans it seems, but the *average* fan
Feels like the pinnacle of creativity https://t.co/5CB3sktSS6
|
2 |
0 |
2 |
597 |
194 |
. |
| Mar 02 |
Problem: Some of the people whose ideas I resonate with/admire most, I also have the most points-of-disagreement with, but if they don’t see the positive tweets and only the disagreements, it looks like I’m just a hater, when the opposite is true. Solution?
|
1 |
0 |
0 |
621 |
257 |
. |
| Feb 26 |
Also referencing this
https://t.co/6CfX7cS7d3
|
0 |
0 |
2 |
899 |
46 |
. |
| Feb 26 |
From recent experience
- There's nothing worse than a freezing bed that you can't fix because it won't connect to wifi.
- There are no better alternatives to @eightsleep from what I can see. The cheaper ones have bad reviews and/or blow air under the covers.
|
0 |
0 |
0 |
685 |
259 |
. |
| Feb 16 |
This is true of every framework. There's a misdirected obsession with the idea that the *framework* needs to be constantly improved. Frameworks should add the *abstractions* that are missing in current web standards, and their only major improvements should be (gracefully) removing those abstractions as it becomes possible to do those things natively
|
4 |
1 |
1 |
840 |
353 |
. |
| Feb 02 |
Another bad faith argument from @cramforce that there's no way to keep a chat app serving ~500 users moderately reliable for less than $90k/year because you have to have a person on call to fix it if it goes down.
The main reason Slack needs people on-call is that they serve the same product to 30 million users. They've had 8 outages in the last 12 months - that's *super* impressive at that scale, but even so, much less reliable than the 6 apps I've been hosting on Cloudron for the last 18 months with zero downtime.
For the "Clearly he's never done this in real life" crowd
- Updates: Handled 100% by Cloudron - all apps are auto updated with new releases.
- Dodgy Deploys: We had one for Gitlab - there was an answer on the forum which required rolling back a version - took me 15 minutes & about 2 hours downtime (still less than Slack in the last year). And I wasn't "on call".
- Backups: Set up in 15 minutes and about 20 clicks with Snapshooter
- Resource Issues: Cloudron tells us when resources are getting low (there's 6 apps on a single VM). It's one click to resize the server with our host.
- Storage Issues: It's €14/month for 300GB of storage with our host - even at 500 users your chat database is not growing beyond that for *years*, and when it does, again it's one click to resize.
What am I missing here? *Why* would you need someone on call for this?
|
3 |
0 |
0 |
621 |
1.4k |
. |
| Jan 27 |
Second Irish creator I've seen this week whose hallmark is boorishness (the other being https://t.co/RruavGQr0K)
And I think I kind of... like it. Is this an Irish creator thing?
|
4 |
0 |
1 |
613 |
180 |
. |
| Jan 23 |
A lot of talk about the countdown until all jobs are taken over by AI - I'm looking for a subscription-accountant service at the minute, and all I want is a fixed monthly price to be able to talk to a real human. Any mention of "AI powered advice" is a no straight away, because I need to trust that there's someone who has/feels accountability if things go wrong.
Maybe I'm shortsighted but I can't see that preference changing much (for me, or others) in the short term, no matter how good the answers get.
|
3 |
0 |
4 |
561 |
509 |
. |
| Jan 20 |
I don't often talk about it because @rails has given me so much, but in my opinion most of what rails does/has done on the frontend doesn't pass this test, and it's why I don't use or recommend using turbo/stimulus/strada etc.
|
1 |
0 |
2 |
690 |
226 |
. |
