JeremyHerrman

Gemini 2.5 Flash-Lite has been my go to for cheap document processing at scale (especially with 50% off batch mode), but they are really boiling the frog with pricing increases with each version:

gemini-2.5-flash-lite: $0.10 input / $0.40 output

gemini-3.1-flash-lite: $0.25 input / $1.50 output

gemini-3.5-flash-lite: $0.30 input / $2.50 output (a 6.25x increase over 2.5!)

Now watch them deprecate Gemini 2.5 Flash-Lite in the coming months...

Another happy user of Handy here!

After seeing so many *subscription based* transcription apps all wrapping *open source models*, finding Handy was a real delight and I'm happy to see the author keep on building!

Haha, yes flex apps definitely had a feel to them!

It's easy to criticize but remember, this was back in the days of supporting IE6 and XHR was still relatively new!

Flex's standard UI library was filled with bluish-gray gradients and verdana :)

Here's an article which has a screenshot with a bunch of controls: https://daverupert.com/2023/02/the-case-for-flex-application...

Comic Chat has a special place in my heart because it inspired my first startup back in 2008, a comic creation web app called Chogger. The site grew to 30K monthly users, mostly K-12 educators who wanted to give their students a fun way to write stories.

The comic creator app itself was adobe flex (flash), actionscript 3.0 (like a typed version of javascript), and I remember spending so many hours getting the balloon tail dragging behavior just right...

one of the teachers made a video overview of how it worked: https://www.youtube.com/watch?v=YKT70TBw1vw

This machine specs reminds me of how awful '90s laptop screens, based on a passive matrix, were. Definitely something I don't miss from that era.

While the 1991 Apple PowerBook 100 did have a passive matrix display, the machine it was based on, the Macintosh Portable from 1989, had a crisp active matrix running at 640×400 (even higher resolution than the compact Macintosh desktops with 512×342).

Interestingly Apple tasked Sony with designing the PowerBook 100 by taking the Macintosh Portable and slimming it down as much as possible. They shaved over 10lbs by moving away from the lead acid battery, dropping the floppy drive, and moving to a passive matrix display.

it's tough for people here to admit, but AI is good at summarizing and that includes commit messages and PR fields

but that does NOT mean you get to skip understanding the code & whipping the AI in line during implementation

my preferred workflow:

- think hard about the problem before touching the keyboard

- back and forth with the AI, swatting down bad implementation ideas and poking holes

- settle on a detailed implementation plan

- let the AI go for however long it needs to (usually minutes, sometimes hours)

- review, iterate, test

- "this looks good, commit in focused chunks and create a PR"

- review commits & PR summary, hand edit for clarity

In large-scale combat operations, victory will depend less on which force fields the most advanced weapons and more on which can sustain combat power under persistent attack.

"force fields" had me reading that sentence entirely wrong

I think it's just their version of "Designed by Apple in California"

people are saying this is benchmark is saturated but all of these have occlusion issues, even sol max.

A skilled human artist wouldn't have both legs in front of the bike, or a single straight line representing both leg's crank arms.

yes, microsoft does hold a trademark for notepad. As litigious as MS is I'm pleasantly surprised they've left notepad++ alone.

Unfortunately anthropic is actively eroding community goodwill in regards to potential IP conflicts:

https://abnormal.ai/blog/abnormal-response-to-anthropic-laws...

https://x.com/steipete/status/2016068265391354181

PSA stop putting bigco IP in your project names!

has clawdbot taught us nothing?

I was surprised with how well grok imagine performed when reviewing fast image generation. It balances quality, speed and cost really well with under $0.02 per 1K image within a few seconds.

My OS 9 battlestation is a G4 tower (Digital Audio) with a Sonnet dual 1.6GHz upgrade, 1.5GB RAM and a nvidia GeForce4 Ti which is one of the best OS 9 graphics cards available.

I'm not sure why folks continue to build services with trademarked names they don't own.

If Anthropic had a problem with Clawdbot they are certainly going to take issue with FablePool.

when I saw those pics I thought: all this effort in the ingredients just to burn them

light blonde pancakes taste the best IMO, but char being carcinogenic seems compelling enough on its own to warrant turning down the heat

I was thinking the same thing. The tuning slide is not what you use while playing, it's the separate slide on the bell side of the trombone for fine adjustment to ensure you're in tune with the rest of the band.

Is AI responsible for the committed code? Should AI be blamed when services go down due to the change?

The answer is absolutely not - the developer is responsible whether the code was AI assisted or not, and the dev's name should be attached to it just like any change.

The OP is right: these are ads, plain and simple, and it's a dark pattern for these companies to have attribution enabled by default

I think of the AVP like Steve Jobs once described: headphones for video [0]

I use it a few hours a day, a few times a week exclusively as a virtual display for my macbook.

Being able to put a huge virtual monitor anywhere is the killer feature and I don't tend to use it for anything else. It's indispensable on flights where laptop lids can't even fully open.

Apple got the interactions just about perfect for AR window management, etc.

Of course the main drawbacks are the weight, size, and overall embarrassment from having a giant face computer strapped to your head.

I don't find it very comfortable for more than a couple hours at time, at least with the solo knit band + the belkin head strap. I'd like to get a dual knit band at some point.

But when this sort of hardware gets miniaturized I think many people will prefer it to hunching over a laptop screen. Just like headphones, wearing glasses while typing at a black screen could signal "Do not disturb, I'm working here"

[0] Steve Jobs on Virtual Reality (D3, 2005) https://www.youtube.com/watch?v=bQECSInWVPY

You're underestimating Apple's meticulous planning, which has only become more intense in the Cook era. Bad feature/UX or not, each one of those decisions was calculated.

Read this ars quote from 2010 [0]:

Apple used the small part—one that is not integral to the device’s functionality—to see if the company was capable or producing a custom design to Apple’s specifications. Typically, manufacturers prefer to have at least two sources for parts, so that a supply problem from one supplier won’t halt manufacturing. Since Liquidmetal is only available from one source, Apple needed to make sure the company could deliver.

For Apple Silicon, there was no way they'd make the switch in one go, so they had to figure out a way to hedge that bet. That's what the TouchBar really was, with all its warts and solutions for problems nobody had.

And as someone else in this thread pointed out, the first custom cellular chip wasn't released with a flagship model - they exclusively paired it with the budget iPhone 16e.

Apple is always calculating and hedging.

[0] https://arstechnica.com/gadgets/2010/08/apple-tested-liquidm...

Apple loves to stealth test new tech in full public view by sneaking it into relatively mundane places, so debuting agentic AI via accessibility is very on brand.

A few other examples:

- The Touch Bar was much more than an OLED strip, it was Apple’s first move in the transition to Apple Silicon on macs. The Apple T1 chip in the 2016 Touch Bar MacBooks was the first solely Apple-designed processor to appear in a Mac and took over several responsibilities away from intel chipsets like power management, fans, sleep/wake, access to the camera & mic, and the secure enclave powering touch ID. Then the T2 added encryption of the SSD, audio management, image processing for the camera, and prevented tampering with the boot process

- The iPhone 3G shipped with a Liquidmetal SIM eject tool, which is made from a strong custom metal alloy which is "practically unbendable by hand unless you want to hurt or cut your fingers." Although Apple hasn’t released anything with the alloy since then, now nearly 20 years later Apple is rumored to be using liquid metal in their upcoming foldable iPhone.

- RealityKit had 3D scanning and a lot of other cool AR capabilities for years which didn’t make sense until the Apple Vision Pro was released.

I bought one of these in 2004 and it really was a slice of the future. The screen was way bigger and higher res than anything I had owned prior, and the ability to jailbreak and run emulators gave it a lot more use than just UMD games.

I still have my original PSP 1000 which worked well until my daughter dropped it a couple weeks ago...

it would be funny if this was meant as a goodwill gesture from anthropic to counter all the recent bad press, only for it to cause even more drama

re: US Politics, I view Apple's gift of the gold & glass trophy to Trump more as a humiliation ritual Cook had to endure so that they can continue to uphold their principles, but with a less adversarial government.

Sure it's gross but it does not necessarily signal an abandonment of values from Apple.

I reject pretty much every point of this article, and I worry that it will lead readers down two dark roads: apathy and secrecy

do you really think bigco is going to steal your vibecoded app just because you used their API? ridiculous. They could already do this before AI with their army of devs.

should you hide all of your ideas until they're perfect and ready for millions of users? we all know this goes against a core tenet of startups which is still true today: launch early and often.

promptfoo/openclaw weren't cloned by openai when they got poplular, they were bought for real $$$

also, regarding this:

2009, I bought a refurbished ThinkPad, installed Xubuntu, and started coding.

you can still do this, even with that same 2009 thinkpad. the hard work is in getting your app out there in front of people, coding is just a small piece of a successful business

totally valid to call out apple for issues, but for most folks already accustomed to apple's walled garden, the grass is not greener on the other side

For instance, I've had a ton of issues with Airpods Max so I moved over to the Sony WH-1000XM6, and they do a much worse job sharing connections between multiple devices to the point where I have to reset them constantly to get sound to reliably play from both my phone & computer.

Same for moving to macOS to linux - the issues on add up to the point where the papercuts from macOS are still worth it.

I'm in the thick of it right now as a serial founder with a 5 and a 1 year old. One thing I'm surprised paulg didn't touch on is how much it can evolve your relationship with your partner (for the better!).

Watching my wife have special moments with our kids fills my heart like nothing else.

Seeing her be an amazing mom is like watching your cofounder take on a completely new role outside of their previous experience and crush it. Except it's even better since you're in love with them and have all these biological/chemical signals to help kick that in high gear.

this site is oozing with early 2000s charm - love it

i also like how they use a through-hole transistor as a scale for their miniature CRTs section

be sure to check out another more modern labor of love for CRTs: crtdatabase.com

agreed that Apple will scrap these, but surely some of these will escape (even in a non-functional state) and with enough effort folks will figure out how to boot them.

I have a few prototype apple devices in my collection. Especially with the sheer number of these AI servers it's just a matter of time before they wind up in public.

For anyone who liked Apple's Xserve lineup, it's very cool to get a peek at these rackmount Apple "advanced AI servers"

I'm excited for these to fall into collectors hands in a decade or two.

Disappointed to see continued increased pricing for 3 Flash (up from $0.30/$2.50 to $0.50/$3.00 for 1M input/output tokens).

I'm more excited to see 3 Flash Lite. Gemini 2.5 Flash Lite needs a lot more steering than regular 2.5 Flash, but it is a very capable model and combined with the 50% batch mode discount it is CHEAP ($0.05/$0.20).