I like the originals. The information density and color contrast is excellent. The “New” has unnecessary white space(padding/margin) and seems to require more scrolling. Looks like aesthetic taste over legibility.
Probably donate half of that to my favorite charities and the other half stashed into responsible investment. Then go back to my 9-5 as usual.
Apps keepshiding any crucial info under 5 slide pull down or some obscure menu in each update. Sometimes I feel immense fear of updating. Most essential apps(bank, utility, medical, transport etc) now feel like demo of some ui framework showing off many ui components. Usability is dead.
I actually like the design. It is simple, information dense and very screenshot-able. I can get everything I need with a single glance. 👌
Very nice pitch and all but it feels like there is a disconnect between the product and the pitch. The overall idea is a frankenstein of twitter+leetcode rank+llm powered code generator+buildinpublic platform+job board. All of these has nothinf to do with OSS.
Are these expense numbers monthly or yearly? If MapGO backend is not utilising >50% CPU at peak load, is hosting the frontend on the same VPS viable?
The thinking smooch one is both amusing and weird at the same time. But definitely looks well done, 10/10 will leave a feedback again 👌
How do you use Vercel and VPS separately?
Thanks for being so honest with the reality. It takes immense courage to come clean about the expenses.
I have seen this for a half a billion product. If the sales and marketing was actually competent, the product would soon go to IPO. Instead it died in startup phase because marketing was more concerned about the product’s status than actual customers who wanted to buy it.
True. The ending was eerily sad and frightening. If I were honest, we see the same game and deception taking place everyday in real world. I still can’t decipher the sheep herd, are they complicit with the pigs or are they simply blind followers.
Is it possible to make the nav bar disappear or stay at the top? It kind of feels odd that it floats above the content like this.
I also find the second one cool. First on feels like those old OSS project landing page using default Jekyll theme. But the first one has healthy information density all in one page, second one is more cool but information is not as accessible/obvious.
It is very upsetting to see my bank account provider and utilities provider announcing proudly about their new AI support agent. I am always very happy to pay extra for human support agent.
Isn’t builder.ai bankrupt and considered dead? archive.is/6EgLH
This week, I managed to read the “Animal Farm” by George Orwell. Very interesting refresher. I definitely found it very creepily parallel to political landscape of late.
“All animals are equal. Some animals are more equal than others.” Quote from the book.
I do not intend to engage in such activities. It will simply regurgitate something from its training data coming from astrology or horoscope page scraped from internet.
That is too generic and will probably be agreed by (and explain) nearly 99%(imho) of all the actual indie devs and solopreneurs here. 🤔
Spending time with family should be priority. Work can wait for tomorrow, but life changes from now to next, hence you may not have the same quality time spending opportunity with family later.
Planning to finally start a lean Instapaper, as mozilla announced shutdown of Pocket soon. Instapaper UI got clunky on mobile and “app” on all devices(esp. my eink reader) is not viable.
The goal of the super-apps(i.e. everything app) is to maximise the time user spends on that app, which means more ad views and more data(user activities). Special apps are impactful but if purpose served is generic, people will tend to go back to super apps.
Their stardust instances are randomly available. Other types are always in supply. I really hate their 100/200Mbps egress though. Regarding customer support, EU operators were always “meh” quality though somehow hetzner can do it well. OVH also has bad customer support but cheaper than scaleway.
Wasn’t it a hot trend back in circa 2013-2016 “app for everything”? Literally everything became an app suddenly(at one time even my landlord had an app!) Very odd time. Also every large tech are moving to everything app, one app for shopping, social, message, community, scheduling, travels etc.
Once interviewed with Contabo. Their engineers did not appear very professional/mature. Also I said that their website looks outdated/weird and they were surprised. But their prices are also affordable.
In EU? I use both Hetzner and NetCup. While netcup’s VPS do not have many distro support, their low tier price and generous data limit is unbeatable. For more global deployment, Scaleway or Vultr.
If you have some plan to monetise. Make basic feature set FOSS. But keep advanced features closed(and optional for running self hosted). If you try to only rely on licensing or selling just managed cloud hosted version, you will probably not get much sales.
No worries. Also you may be able to improve overall experience by using multi-threaded range based model download(didn’t check if hf api supports it). Downloading large models is headache on bad residential connections. If Hf supports range, then you can also support resume partial downloads.
Bsky is very limited hence hard to explain well. Also why are you saving/loading token via API calls is strange to me. But then again, looks decent, needs some cleanup and work. But Getting something working is highest priority always. 👌
It is fine for local. Some quick wins: do not modify .env file in user’s home dir, do not use timestamp as threadid, do not use regex to cleanup text, inconsistent exception handling, cache api calls, check if model already exist locally, use basic patterns to set values instead of many if/else.
Interesting. From brief skim it looks verbose. I see some obvious concerning bits at some places but overall looks decent for Proof-of-Concept. Thanks for sharing.
