I’ve been paying $125/mo (promo pricing for the $250/mo tier) for Google’s top-tier “AI Ultra” subscription for about a month.
I signed up because I’m a power user. I manage large repos, build my own MCP tools, and live in the terminal. I assumed “Ultra” meant “Pro Developer.”
I was wrong. I downgraded to the $20/mo plan today. Here’s why other developers should probably avoid the top tier.
The “Double Pay” paradox
The subscription gives you:
The Web UI (Gemini Advanced)
30TB of storage
It gives you:
Zero API credits
If you use modern AI coding tools like Aider, Cursor, or your own CLI scripts, you still pay per-token for the API.
My bill:
$125/mo for the subscription
~$50/mo for API usage
Reality:
I was effectively donating $125/mo to Google for a Web UI I barely use, because I work in the CLI.
The “official” CLI is broken
I tried to use the official gemini-cli to avoid API costs. It’s unusable for agentic work.
There’s a known, persistent bug where the CLI crashes with a 400 Bad Request error whenever the model attempts parallel tool calling (a standard model feature).
Error:
“Please ensure that the number of function response parts is equal…”
The pricing only makes sense if you look at Google’s internal structure:
Google One wants to sell storage (30TB bundle)
Google Cloud wants to sell APIs (Vertex)
DeepMind wants to ship models
“AI Ultra” is a Franken-product stitched together by the storage team. There’s no billing concept for “individual developer who needs API credits,” so instead you’re forced to buy a data center’s worth of hard drives.
The feedback black hole
There’s no way to get this feedback to someone who can act on it. You can’t file a ticket that says “your pricing model ignores developers.” You just shout into a generic feedback form.
The fix: unbundle the intelligence
What’s needed is a real Developer Tier:
Give me: high API rate limits + API credits
Charge me: fairly for what I use
Until that exists, I’ve downgraded to the $20/mo plan and moved the rest of my AI budget to Anthropic.
The number of times I've used GCP products and they barely work, really tells me normal engineers don't test the products only wizards that think things are functional.
Hinge Health | All Levels, Engineering & Product | Remote USA/Canada (office hubs also) | https://www.hingehealth.com/
Hinge Health is the leading provider of digital musculoskeletal therapy, basically we provide an app that enables exercise to improve conditions such as knee or back pain (it works, see https://pubmed.ncbi.nlm.nih.gov/32208358/). We change people's lives for the better every day and we get to read their feedback.
We’re looking to hire full-time developers, across our stack, technologies we use are typescript, react, react-native, ruby on rails, node.js, python, and more -- ultimately we want to use the best tools for the job in a microservices environment.
We’re thriving as a business, valued at ~$6 billion during the last round of funding. Additionally we have SRE. QA engineer, hardware, embedded software eng, health coach, management, product and design positions available.
Hinge Health | All Levels, Engineering & Product | SF, South Bay, LA CA, Portland OR, Boulder/Denver CO, Austin TX, Minneapolis MN, Chicago IL, New York NY (WFH/REMOTE during pandemic) | https://www.hingehealth.com/
Hinge Health is the leading provider of digital musculoskeletal therapy, basically we provide an app that enables exercise to improve conditions such as knee or back pain (it works, see https://pubmed.ncbi.nlm.nih.gov/32208358/). We change people's lives for the better every day and we get to read their feedback.
We’re looking to hire full-time developers, across our stack, technologies we use are typescript, react, react-native, ruby on rails, node.js, python, and more -- ultimately we want to use the best tools for the job in a microservices environment.
We’re thriving as a business, valued at ~$6 billion during the last round of funding. Additionally we have SRE. QA engineer, hardware, embedded software eng, health coach, management, product and design positions available.
Hinge Health | All Levels, Engineering & Product | SF, South Bay, LA CA, Portland OR, Boulder/Denver CO, Austin TX, Minneapolis MN, Chicago IL, New York NY (WFH/REMOTE during pandemic) | https://www.hingehealth.com/
Hinge Health is the leading provider of digital musculoskeletal therapy, basically we provide an app that enables exercise to improve conditions such as knee or back pain (it works, see https://pubmed.ncbi.nlm.nih.gov/32208358/). We change people's lives for the better every day and we get to read their feedback.
We’re looking to hire full-time developers, across our stack, technologies we use are typescript, react, react-native, ruby on rails, node.js, python, and more -- ultimately we want to use the best tools for the job in a microservices environment.
We’re thriving as a business, valued at ~$6 billion during the last round of funding. Additionally we have SRE. QA engineer, hardware, embedded software eng, health coach, management, product and design positions available.
Hinge Health | All Levels, Engineering & Product | SF, South Bay, LA CA, Portland OR, Boulder/Denver CO, Austin TX, Minneapolis MN, Chicago IL, New York NY (WFH/REMOTE during pandemic) | https://www.hingehealth.com/
Hinge Health is the leading provider of digital musculoskeletal therapy, basically we provide an app that enables exercise to improve conditions such as knee or back pain (it works, see https://pubmed.ncbi.nlm.nih.gov/32208358/). We change people's lives for the better every day and we get to read their feedback.
We’re looking to hire full-time developers, across our stack, technologies we use are typescript, react, react-native, ruby on rails, node.js, python, and more -- ultimately we want to use the best tools for the job in a microservices environment.
We’re thriving as a business, valued at ~$6 billion during the last round of funding. Additionally we have SRE. QA engineer, hardware, embedded software eng, health coach, management, product and design positions available.
PayPal refuses to let heirs access, or even know if there is a balance on accounts after people die, regardless of death certificates. I wonder how much money is being held by this tactic?
Hinge Health | All Levels, Engineering & Product | SF, South Bay, LA CA, Portland OR, Boulder/Denver CO, Austin TX, Minneapolis MN, Chicago IL, New York NY (WFH/REMOTE during pandemic) | https://www.hingehealth.com/
Hinge Health is the leading provider of digital musculoskeletal therapy, basically we provide an app that enables exercise to improve conditions such as knee or back pain (it works, see https://pubmed.ncbi.nlm.nih.gov/32208358/). We change people's lives for the better everyday and we get to read their feedback.
We’re looking to hire full-time developers, across our stack, technologies we use are typescript, react, react-native, ruby on rails, node.js, python, and more -- ultimately we want to use the best tools for the job in a microservices environment.
We’re thriving as a business, and received $300M series D funding in January, at a $3 billion valuation.Additionally we have SRE. QA engineer, hardware, embedded software eng, health coach, management, product and design positions available.
Hinge Health | All Levels, Engineering & Product | SF, South Bay, LA CA, Portland OR, Boulder/Denver CO, Austin TX, Minneapolis MN, Chicago IL, New York NY (WFH during pandemic) | https://www.hingehealth.com/
Hinge Health is the leading provider of digital musculoskeletal therapy, basically we provide an app that allows people to use exercise to improve conditions such as knee or back pain (it works, see https://pubmed.ncbi.nlm.nih.gov/32208358/). We change people's lives for the better everyday and we get to read their feedback.
We’re looking to hire full-time developers, across our stack, technologies we use are typescript, react, react-native, ruby on rails, node.js, python, and more -- ultimately we want to use the best tools for the job in a microservices environment.
We’re thriving as a business despite the pandemic, and received $300M series D funding in January, at a $3 billion valuation.
Additionally we have SRE. QA engineer, hardware, embedded software eng, health coach, management, product and design positions available.
> Although the first hardware run will be entirely $140 / 8GiB systems, lower-cost variants with less RAM are expected in following releases.
> The initial pilot run of BeagleV will use the Vision DSP hardware as a graphics processor, allowing a full graphical desktop environment under Fedora. Following hardware runs will include an unspecified model of Imagine GPU as well.
Sounds like a direct competitor to the Raspberry Pi. I don't know if the Imagine GPU planned for the next iteration is playing catch-up or leapfrog. The Arstechnica article links to SiFive creates global network of RISC-V startups [1] which I think demonstrates that SiFive is strategically leveraging or responding to the geopolitics surrounding Chinese technology.
Imagination GPU :( Notorious for being hard to support in open source. I'm not even sure there was a single free driver for those.
That likely means those devices are going to be stuck on an outdated kernel, unless Imagination steps in and provides ongoing binary support for newer kernels for their GPUs like x86 GPU manufacturers do. However, this being RISC-V with 2 existing devices total, I don't count on it.
Except for cost... which has been a problem for the BeagleBoard line of SBC's since the beginning. They actually predated the original Raspberry Pi by a couple of years but when the Pi came in at ~25% of the cost, they caught up and overtook the BeagleBoard in popularity fast. The BeagleV looks interesting from an early adopter standpoint but the hobbyist market will probably standardize around whatever decent RISC-V board comes in at sub-$50 first.
To me, they seem to serve different markets. The various BeagleBoards have more industrial specs like a wider operating temperature range, on-board EMMC, etc. Also, the pair of PRU's make them useful for things where more precise timing is important.
Regarding the BBB/BBG, in the last 3-5 years the RPis have gotten significantly faster (RPi3 & 4) and gone 64-bit whereas the BBB & BBG haven't changed much (aside from a bit more eMMC and a very minor CPU bump) since they were launched. These days the 1GHZ 32-bit AM3358 (BBB RevC) is comparatively much slower and with only 512MB RAM, that's a lot less than a stock RPi 4.
Having said that, the BBBs are a great device! They're rock solid and have far better I/O options than the RPi: 4 UARTS, multiple I2C, SPI & CAN buses, EHRPWM, a ton of GPIO, 2x PRU processors, LCD driver, both USB and USB Gadget, oh and of course, the onboard eMMC is great compared to booting from an SD.
>>*SiFive is strategically leveraging or responding to the geopolitics surrounding Chinese technology.*
interesting subtext.
We should all be debating HW WRT the fact that you can't name a single device (aside from weapons) that does not contain a single-non-chinese-manufactured component...
every phone or machine is almost 100% chinese built.
"designed by apple in cupertino california" (but made with slave labor from congo, china and other countries)
And we already know about all the backdoors both China and the US do...
FFS we have known about Eschelon since the 70s - the carnivor, room 641A, etc... etc....
Indeed, the geopolitics works both ways. I think the Chinese are looking at RISC-V as a safe-guard against American embargoes of the kind that killed/maimed HiSilicon, the non-Chinese nation-states are looking for full transparency of silicon design, and the manufacturers want full access to a truly global market that includes China. I'm not sure that SiFive RISC-V designs can be competitive with ARM/x64 in the short-term but the geopolitics creates a potential niche.
I think you are conflating assembly and manufacturing. TSMC is Taiwanese and Samsung is South Korean. Personally I'd prefer that all nation-states and their security organizations followed the Golden Rule and promoted free trade rather than protectionism.
> made with slave labor from congo, china and other countries
I don't equate low-wage manufacturing/assembly with exploitation and certainly not slavery but I understand that this is a common metaphor. Contemporary slavery [1] is a real thing and, until I see contrary evidence, I'm assuming it makes zero contribution to high-tech assembly or manufacturing.
We have min wage in the US, and even that is not equal across all states - we do not have UBI or universal health care, we have shitty industries, such as insurance (forced hedge funds) and in general, we have brainwashing people to accept it as normal.
fuck that.
YC even wants you to think that all VC is all-truistic NOPE.
I signed up because I’m a power user. I manage large repos, build my own MCP tools, and live in the terminal. I assumed “Ultra” meant “Pro Developer.”
I was wrong. I downgraded to the $20/mo plan today. Here’s why other developers should probably avoid the top tier.
The “Double Pay” paradox
The subscription gives you:
The Web UI (Gemini Advanced)
30TB of storage
It gives you:
Zero API credits
If you use modern AI coding tools like Aider, Cursor, or your own CLI scripts, you still pay per-token for the API.
My bill:
$125/mo for the subscription
~$50/mo for API usage
Reality: I was effectively donating $125/mo to Google for a Web UI I barely use, because I work in the CLI.
The “official” CLI is broken
I tried to use the official gemini-cli to avoid API costs. It’s unusable for agentic work.
There’s a known, persistent bug where the CLI crashes with a 400 Bad Request error whenever the model attempts parallel tool calling (a standard model feature).
Error: “Please ensure that the number of function response parts is equal…”
These issues have been open for months:
Broken auth: https://github.com/google-gemini/gemini-cli/issues/11925
Parallel tool crash: https://github.com/google-gemini/gemini-cli/issues/13292
Shipping the org chart
The pricing only makes sense if you look at Google’s internal structure:
Google One wants to sell storage (30TB bundle)
Google Cloud wants to sell APIs (Vertex)
DeepMind wants to ship models
“AI Ultra” is a Franken-product stitched together by the storage team. There’s no billing concept for “individual developer who needs API credits,” so instead you’re forced to buy a data center’s worth of hard drives.
The feedback black hole
There’s no way to get this feedback to someone who can act on it. You can’t file a ticket that says “your pricing model ignores developers.” You just shout into a generic feedback form.
The fix: unbundle the intelligence
What’s needed is a real Developer Tier:
Give me: high API rate limits + API credits
Charge me: fairly for what I use
Until that exists, I’ve downgraded to the $20/mo plan and moved the rest of my AI budget to Anthropic.
reply