• Latest
  • Trending
  • All
Answer card: Google begins removing Google Assistant from Android phones, tablets, Wear OS watches, compatible headphones and projected Android Auto on 4 September 2026, a rollout expected to take a few weeks, after which the device cannot switch back to Assistant.

Google Assistant shuts down September 4: what you lose

5 August 2026
Answer card stating that Qwen-Image-2.1, released on 20 September 2026, ships open weights with a 7 billion parameter diffusion transformer, a Qwen3-VL 8B text encoder and an RGBA VAE totalling about 33 gigabytes in BF16, under the Qwen Research License that limits use to research or evaluation and requires a separate commercial licence, unlike the Apache 2.0 licence of Qwen-Image 1.0.

Qwen-Image-2.1 brings the weights back, but not the Apache licence

21 September 2026
Answer card stating that Ternary Bonsai 2 27B, released by PrismML on 17 September 2026 under Apache 2.0, packs Qwen3.8 27B into 5.95 gigabytes at 1.72 bits per weight, keeps 98.2 percent of the 14-benchmark average, about 75 percent on SWE-bench Verified and Terminal-Bench 2.1, and needs PrismML's llama.cpp fork to run.

Does Bonsai 2 27B really keep 98% of Qwen3.8 in 5.95 GB?

20 September 2026
Answer card stating that Jev 1.13 from TypeSafe AI is a decision model in early access since 15 September 2026 that returns typed probabilities instead of text, priced at 42 dollars per billion input tokens with output tokens free, answering in 70 to 500 milliseconds, with a 64K token request budget, text input only, and a documented list of things it does badly, including counting and dates.

Jev 1.13 bills $42 a billion tokens, and it can’t count

19 September 2026
Answer card stating that Qwen3.8-Omni-Flash launched on 17 September 2026 as an API only model on Alibaba Cloud Model Studio, taking text, images, audio and video in a 1M token context and returning text only, priced at 0.15 dollars per million input tokens for every modality and 0.47 dollars per million output tokens in the international regions, with no open weights published and the Qwen-Live Harness GitHub repository returning 404.

Qwen3.8-Omni-Flash bills audio at $0.15 and ships no weights

18 September 2026
Answer card stating that on 15 September 2026 AWS said it is unable to restore access to resources and data hosted exclusively in the Middle East Bahrain region me-south-1 and in the mec1-az2 zone of the UAE region, because the damage spanned multiple Availability Zones and exceeded what multi-AZ services are designed to withstand.

AWS can’t restore me-south-1, six months after the drone strikes

17 September 2026
Answer card stating that Google released Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking on 15 September 2026 at 3 dollars per million audio input tokens and 12 dollars out, that the thinking model requires asynchronous tools, and that Artificial Analysis scores it 82.6 on its Speech to Speech Quality Index.

Gemini 3.8 Live Extended Thinking rejects any tool that blocks

16 September 2026
Answer card summarising the Atria Dawn Preview release: 744B GLM-5.2 base, MIT licence, 1.5 TB BF16 and 756 GB FP8 checkpoints, 256K context, top on five of sixteen benchmark rows and trailing on SWE-bench Pro.

Atria Dawn Preview is 744B under MIT, and the BF16 weighs 1.5 TB

15 September 2026
Answer card stating that OpenAI released the Agents API in public beta on 10 September 2026 with no separate fee, billed through model tokens, tool calls and hosted sandbox time, with a choice of OpenAI hosted, self hosted or partner sandboxes, US only data residency and no Zero Data Retention support.

OpenAI’s Agents API has no fee, no ZDR and a one hour sandbox clock

14 September 2026
Answer card: Sakana Fugu Max at $2 and $6 per million tokens, Fugu Ultra v2 unchanged at $5 and $30, and Sakana saying Ultra v2 scores without Fable 5 or GPT-6 Astra in its pool.

Fugu Max costs $2 and $6 while Fugu Ultra v2 runs without Fable 5

13 September 2026
Answer card stating that DeepSeek released DeepSeek-V4.1-Flash on 10 September 2026 as a 552 billion parameter mixture of experts model with a new causal encoder decoder architecture that activates 8 billion parameters on input and 16 billion on output, with native vision, a one million token context and MIT licensed weights, that the API model name is now deepseek-flash at 0.15 dollars per million input tokens and 0.60 dollars per million output tokens off peak, and that DeepSeek announced V4 Pro would be routed to V4.1-Flash from 14 September and reversed that on 11 September.

DeepSeek V4.1-Flash arrived, and the V4 Pro retirement lasted a day

12 September 2026
Answer card stating that Cognition released SWE-2 on 10 September 2026, a coding model post-trained from Kimi K3, scoring 50.0 percent on FrontierCode 1.1 Main against 50.9 percent for Claude Fable 5.1 and 27.3 percent on Terminal-Bench 4 against 55.8 percent, available only inside Devin.

SWE-2 trails Fable 5.1 by one point, and by 28 on Terminal-Bench 4

11 September 2026
Answer card for Meta Muse, free to 100 million tokens a week then $20 a month, launched 8 September 2026 for United States adults only, running in a dedicated per user virtual machine.

Does Meta Muse do enough to earn your inbox and a card on file?

9 September 2026
  • About
  • Contact
  • Privacy
  • Legal
Tuesday, September 22, 2026
  • Login
Packet Nebula
  • Home
  • Articles
    • Security
    • Network
    • Dev
    • Sysadmin
    • SEO
    • Email & DNS
  • Tools
    • Network tools: free, fast, no signup
    • Security tools: free, fast, no signup
    • Developer tools: free, fast, no signup
    • Sysadmin tools: free, fast, no signup
    • SEO tools: free, fast, no signup
    • Email & DNS tools: free, fast, no signup
  • Download
  • About
No Result
View All Result
Packet Nebula
No Result
View All Result
Home Dev

Google Assistant shuts down September 4: what you lose

by stephane
5 August 2026
in Dev
0
Answer card: Google begins removing Google Assistant from Android phones, tablets, Wear OS watches, compatible headphones and projected Android Auto on 4 September 2026, a rollout expected to take a few weeks, after which the device cannot switch back to Assistant.
491
SHARES
1.4k
VIEWS
Share on FacebookShare on Twitter

Check your inbox. If there is an Android phone in your pocket, Google sent you a notice this week saying the assistant on it is being discontinued. Removal starts 4 September 2026, it reaches phones, tablets, Wear OS watches and any headphones paired to that phone, and once it lands on a device you cannot go back. That last clause is the one to read twice. Google''s own wording is that you will no longer be able to use or switch back to Google Assistant on your phone, tablet, or paired devices. Cars with Google built in keep Assistant past that date. Android Auto projected from your phone does not.

The short answer

Google is retiring Google Assistant on mobile. Removal starts 4 September 2026 and rolls out over a few weeks, covering Android phones and tablets plus the Wear OS watches, compatible headphones and projected Android Auto sessions attached to them. Cars with Google built in are exempt for now, and TVs and speakers are a separate announcement. The migration is one way. Gemini still has holes where Assistant had features.

Sept 4removal begins
One wayno switching back
2016when Assistant launched
Answer card: Google begins removing Google Assistant from Android phones, tablets, Wear OS watches, compatible headphones and projected Android Auto on 4 September 2026, a rollout expected to take a few weeks, after which the device cannot switch back to Assistant.
One start date, a rollout of unknown length, and a door that only opens outward.

The email, and the sentence inside it

Google did not blog this. It mailed it, which is why the date turned up in user inboxes on 4 August before it turned up anywhere official looking.

The notice opens by saying the Google Assistant on your mobile device is being discontinued and that Gemini is now the assistant experience on Android. Fine. Everyone saw that coming since Gemini started answering the power button long press. The operative paragraph is the next one: “We will begin removing access to Google Assistant starting on September 4, and expect that this process may take a few weeks to reach everyone. Once availability is removed, you will no longer be able to use or switch back to Google Assistant on your phone, tablet, or paired devices.”

Read the second sentence again. Right now, on most Android phones, the assistant choice is a setting. You can put Gemini on, decide you hate what it did to your alarms, and put Assistant back. That reversibility is what has made the last eighteen months of this migration tolerable.

It ends on 4 September. Not the ability to use Gemini, the ability to refuse it.

Honestly, we think the staggered rollout is the more annoying half of this. “A few weeks to reach everyone” means you cannot plan a cutover. Your phone might flip on the 4th and your spouse’s identical phone on the 22nd, same account, same model, no notification when it happens.

What loses it, what keeps it

The email is unusually precise about surfaces, so it is worth being precise back.

Going: Android phones and tablets that meet Gemini’s minimum requirements, in regions where Gemini ships. Then everything downstream of that phone. Watches running Wear OS. Headphones and earbuds compatible with Gemini. Vehicles running Android Auto projected from the phone.

Diagram comparing the two groups: Android phones and tablets, Wear OS watches, compatible headphones and projected Android Auto lose Google Assistant from 4 September 2026, while cars with Google built in, Google TV devices, Google Home speakers and smart displays keep it for now under separate plans.
The split is not about the brand on the device. It is about whether your phone is the thing running the assistant.

Staying, at least for now: cars with Google built in, which Google names outright as continuing to function beyond 4 September 2026. Google TV, Google Home speakers and smart displays are not in this announcement at all, and Google points to separate plans for bringing Gemini to them.

That Android Auto line catches people. Two cars can show the same interface and land on opposite sides of this, because one is projecting from a handset and the other is running its own Google account in the dashboard. If you cannot remember which yours is, the giveaway is whether the car still does assistant things with the phone unplugged and out of the vehicle.

The gap list nobody put in the email

Here is where the announcement gets quiet. It promises a smooth transition and a more capable assistant. It does not enumerate what Gemini currently cannot do.

Reporting on the migration has been tracking those gaps for months, and the list is not short. Podcasts, news and radio stations and third party music providers do not have the same hooks. Routines are the ugly one: Gemini does not support every starter or action, scheduled time triggers and some media actions among the holes, and routines fired from Clock app alarms have been reported as simply not running. Reminders and task creation have been flagged as missing. So has interpreter mode. Google Home device control works by voice but not by typing.

Checklist of what is confirmed about the Google Assistant to Gemini migration and what remains unresolved, covering the 4 September 2026 start date, the one way nature of the change, exempt device classes, and reported feature gaps in Gemini around routines, reminders, media services and interpreter mode.
Four things Google put in writing, four things it did not.

Most of these read like unfinished plumbing rather than deliberate removals, and some will probably be quietly fixed before September. I might be wrong about that. The point is that Google has published no dated commitment on any single one of them, and the deadline is dated. Those two facts are not symmetrical, and you are the one absorbing the difference.

If your household automation leans on Assistant routines, that is the thing to test now, while you can still fall back.

What to do in the next four weeks

Nothing dramatic. Just stop treating the switch as reversible, because in a month it is not.

Flip to Gemini deliberately, on your own schedule, and run your actual daily commands through it. The alarm you set by voice every night. The lights. Whatever the watch does on your wrist without you looking at it. You will find the broken ones faster by using it for a week than by reading any feature comparison, including this one.

Then decide what needs a non-assistant backup. A routine that matters becomes a Clock alarm plus a Home automation. A voice reminder becomes an app you actually open. Boring, but it survives the next assistant too, and going by the last decade there will be a next assistant.

For the fleet managers reading this: there is no enterprise control here. No policy, no deferral, no staged ring. The rollout reaches devices when it reaches them. If you have kiosk tablets or shared handsets with voice workflows on them, that workflow changes underneath you sometime after 4 September and you will not be told which day.

We covered what changed in Gemini 3.6 Flash when the model side of this moved, and the new Siri in the iOS 27 public beta if you want to see the other big assistant rewrite happening on the same calendar.

Ten years is a decent run for a product. Assistant shipped in 2016 and it is being switched off by email in 2026, which is roughly how these things go now.

Sources

Google’s notice reached users by email rather than a blog post. The full text is reproduced by Droid Life, and the date and affected surfaces were reported independently by 9to5Google, Android Authority and Engadget. The running list of features Gemini has not yet matched comes from Android Police. Where the reporting disagreed, we went with the wording in Google’s own email.

Frequently asked questions

What is the exact date Google Assistant shuts down?

Google begins removing access on 4 September 2026. That is a start date, not a finish line. The email says the process may take a few weeks to reach everyone, so two phones on the same account can flip on different days. There is no published end date for the rollout and no way to request it early or late.

Can I switch back to Google Assistant afterwards?

No. This is the part people keep missing. Google wrote that once availability is removed, you will no longer be able to use or switch back to Google Assistant on your phone, tablet, or paired devices. Today the Assistant settings screen still offers a toggle between the two. After the change reaches your device that toggle is gone, and it is gone for the watch and the earbuds attached to it as well.

Which devices keep Google Assistant?

Cars with Google built in are named explicitly in Google's email as continuing to work beyond 4 September 2026. Google TV devices, Google Home speakers and smart displays are outside this announcement entirely, and Google says it will bring Gemini to them under separate plans. Android phones that do not meet Gemini's minimum requirements, or that sit in a region where Gemini is not offered, are also not part of this wave.

What can Gemini not do that Assistant could?

Enough to matter if you built habits on it. Podcasts, news and radio stations and third party music providers are not wired up the same way. Routines are partial, with some starters and actions unsupported, and routines attached to Clock app alarms have been reported as broken. Setting reminders and creating tasks, interpreter mode, and typed rather than spoken control of Google Home devices have all been flagged as missing. Google has published no dated commitment to restore any specific one.

Does this affect Android Auto?

It affects the projected kind. If Android Auto runs off your phone and mirrors to the head unit, that stack follows the phone and moves to Gemini. If the car itself runs Google built in, with its own account and its own Assistant, Google says it keeps working past 4 September 2026. Same badge on the dashboard for a lot of drivers, two different outcomes.

Tags: androidassistantbig-techgeminigooglenews
Share196Tweet123
stephane

stephane

  • Trending
  • Comments
  • Latest
Answer card: Proton Lumo 2.0 is private by policy, not by locality. Saved history is locked so even Proton cannot read it, but the prompt is decrypted on a Proton EU server to answer it, then forgotten.

Proton Lumo 2.0 review: how private is it, really?

3 September 2026
The Agentic Coding section of the official Hy4 preview benchmark appendix published by Tencent, a table comparing Hy3 and Hy4 preview against DeepSeek V4 Pro 0813, Qwen 3.8 Max, GLM 5.3, Kimi K3, GPT 5.6 Sol and Claude Opus 5 across SWE-bench Multilingual, SWE-bench Pro, DeepSWE, three SWE Atlas tasks, SWE-Marathon, Terminal-Bench 2.1, NL2Repo-Bench, CyberGym, ProgramBench, PostTrainBench and Harbor-Index.

Tencent’s 770B Hy4 tops one benchmark row in 46

3 September 2026
Answer card: Qwen 3.7 Max is API-only and cannot run locally yet; the open Qwen models (Qwen 3.6 27B, qwen3:8b to 32b) run offline via Ollama.

Qwen 3.7 local: what you can actually run offline

22 June 2026
Answer card: JWTs are not encrypted, anyone can read them; the signature proves who issued the token, not who may read it.

Are JWTs encrypted? No, and the difference will bite you

0
Answer card: a random 8 character password falls in under 2 hours offline, while 16 random characters hold for 1.4 trillion years at the same speed.

How long does it take to crack a password in 2026?

0
Answer card: three DNS records decide if your mail lands or bounces; SPF lists allowed senders, DKIM signs messages, DMARC sets the failure policy.

SPF, DKIM and DMARC explained: the records your email needs

0
Answer card stating that Qwen-Image-2.1, released on 20 September 2026, ships open weights with a 7 billion parameter diffusion transformer, a Qwen3-VL 8B text encoder and an RGBA VAE totalling about 33 gigabytes in BF16, under the Qwen Research License that limits use to research or evaluation and requires a separate commercial licence, unlike the Apache 2.0 licence of Qwen-Image 1.0.

Qwen-Image-2.1 brings the weights back, but not the Apache licence

21 September 2026
Answer card stating that Ternary Bonsai 2 27B, released by PrismML on 17 September 2026 under Apache 2.0, packs Qwen3.8 27B into 5.95 gigabytes at 1.72 bits per weight, keeps 98.2 percent of the 14-benchmark average, about 75 percent on SWE-bench Verified and Terminal-Bench 2.1, and needs PrismML's llama.cpp fork to run.

Does Bonsai 2 27B really keep 98% of Qwen3.8 in 5.95 GB?

20 September 2026
Answer card stating that Jev 1.13 from TypeSafe AI is a decision model in early access since 15 September 2026 that returns typed probabilities instead of text, priced at 42 dollars per billion input tokens with output tokens free, answering in 70 to 500 milliseconds, with a 64K token request budget, text input only, and a documented list of things it does badly, including counting and dates.

Jev 1.13 bills $42 a billion tokens, and it can’t count

19 September 2026
  • About
  • Contact
  • Privacy
  • Legal

Copyright © 2026 Stephane Cardon.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In
No Result
View All Result
  • Home
  • Articles
    • Security
    • Network
    • Dev
    • Sysadmin
    • SEO
    • Email & DNS
  • Tools
    • Network tools: free, fast, no signup
    • Security tools: free, fast, no signup
    • Developer tools: free, fast, no signup
    • Sysadmin tools: free, fast, no signup
    • SEO tools: free, fast, no signup
    • Email & DNS tools: free, fast, no signup
  • Download
  • About

Copyright © 2026 Stephane Cardon.