zhyder

AI dramatically reduces the cost of writing code (especially when you have a reference), so the scale between native and cross-platform is going to tilt more towards native now compared to before.

But I'm curious what their update will be in a year or two. Because these costs don't reduce as dramatically with AI (unless you fully give up control and vibe code it):

1. Reading code

2. Manually testing code

A big part of this announcement does seem to be _delegation_ in the background; they give the example of web search but that could be any tool. I haven't tried it yet either but sounds like they've found a reasonable UX that mixes that sort of high-and-variable latency tool calling (potentially with agentic loops) with a continuously speaking live voice.

Aside from the issue of platform owners (Apple, Google, Microsoft) offering storage sync as an integrated feature, which others have pointed out, the other reason growth is limited is that filesystem storage and sync thereof has become less critical over time. Our apps increasingly do cloud-native walled-garden storage: docs in Google Docs / Notion / Confluence, mocks in Figma, etc. And code was cloud-native with version control much earlier.

I need sync for just photos on my phone (which Apple or Google are better for), and a small number of esigned PDFs and tax documents (for which any provider's free tier suffices).

Dropbox solved a problem of the 2010s.

Looks like the best display you can get in laptops at this price: 2408x1506 resolution, 500 nits, antireflective coating (!). And bonus points for no silly notch.

I guess it could warn about it but the VM sandbox is the best part of Cowork. The sandbox itself is necessary to balance the power you get with generating code (that's hidden-to-user) with the security you need for non-technical users. I'd go even further and make user grant host filesystem access only to specific folders, and warn about anything with write access: can think of lots of easy-to-use UIs for this.

Model card: https://deepmind.google/models/model-cards/gemini-3-1-flash-...

Pretty close to Gemini 3 Pro Image (aka Nano Banana Pro) in most benchmarks, even without thinking+search, and even exceeding it in 2 most important ones of 'Overall Preference' and 'Visual Quality'. I'm excited about the big jump in Infographics/Factuality (even without thinking+search; I'm surprised that text+image search grounding doesn't make an even bigger dent).

Agree, can't wait for updates to the diffusion model.

Could be useful for planning too, given its tendency to think big picture first. Even if it's just an additional subagent to double-check with an "off the top off your head" or "don't think, share first thought" type of question. More generally would like to see how sequencing autoregressive thinking with diffusion over multiple steps might help with better overall thinking.

Surprisingly big jump in ARC-AGI-2 from 31% to 77%, guess there's some RLHF focused on the benchmark given it was previously far behind the competition and is now ahead.

Apart from that, the usual predictable gains in coding. Still is a great sweet-spot for performance, speed and cost. Need to hack Claude Code to use their agentic logic+prompts but use Gemini models.

I wish Google also updated Flash-lite to 3.0+, would like to use that for the Explore subagent (which Claude Code uses Haiku for). These subagents seem to be Claude Code's strength over Gemini CLI, which still has them only in experimental mode and doesn't have read-only ones like Explore.

"the value of a human eyeball" / attention is and always will be the limited resource. But I wish the way the economy worked wasn't that attention is sold for money, which makes money the moat, and sets a floor on how low-priced things can get for customers too. Is this really the best the economy can do? Or is it possible to have a fair LLM-based search engine that matches customer need description with stated product descriptions from providers (while weighing customer reviews, etc)?

Hmm the whole point of checkpoints seems to be to reduce token waste by saving repeat thinking work. But wouldn't trying to pull N checkpoints into context of the N+1 task be MUCH more expensive? It's at odds with the current practice of clearing context regularly to save on input tokens. Even subagents (which I think are the real superpower that Claude Code has over Gemini CLI for now) by their nature get spawned with fresh near-empty context.

Token costs aside, arguably fresh context is also better at problem solving. When it was just me coding by hand, I didn't save all my intermediate thinking work anywhere: instead thinking afresh when a similar problem came up later helped in coming up with better solutions. I did occasionally save my thinking in design docs, but the equivalent to that is CLAUDE.md and similar human-reviewed markdown saved at explicit -umm- checkpoints.

So 2.5x the speed at 6x the price [1].

Quite a premium for speed. Especially when Gemini 3 Pro is 1.8x the tokens/sec speed (of regular-speed Opus 4.6) at 0.45x the price [2]. Though it's worse at coding, and Gemini CLI doesn't have the agentic strength of Claude Code, yet.

[1] - https://x.com/claudeai/status/2020207322124132504 [2] - https://artificialanalysis.ai/leaderboards/models

Love it. Wonder if it's viable for citizen journalism in warzones and areas of civil unrest, with the larger size of photos (and short videos), given the inherently slow transfer rates and battery life implications of going thru multiple hops before Internet-exiting the area that's otherwise Internet-offline. What's the back-of-the-envelope math here on viable bandwidth?

Wifi obviously has higher bandwidth, but I guess it isn't viable as a mesh, or is there any trick with turning on/off hotspots on phones dynamically that'd make it viable? (Afaik older phones made you pick between being a hotspot or being a regular wifi client, but at least some newer ones seem to allow both simultaneously.)

I'm definitely hoping for a future with wider support for C2PA (content credentials on images) on phone cameras to make these photos power citizen journalism. So far Samsung S25 and Pixel 10 support C2PA in the camera hardware: need other phone makers (especially Apple) to get on board already... if you're an iPhone user, please help yell at Apple support etc!

Aside: I registered a domain and plan to build a citizen journalism news feed for such photos (and uncut videos). I see it as the antidote to Instagram et al's feeds that're full of AI slop (and plenty of fakery even before AI-generated imagery got big). And it's essential to truth, democracy and ultimately (maybe I'm too idealistic here) peace. Aside to the aside: wish some of us techies banded together to build "peace tech" as a new sector in tech, DM if interested in brainstorming or working together.

Sounds like antirez, simonw, et al are still advocating reviewing the code output of these agents for now. But presumably soon (within months?) the agents will be good enough such that line-by-line review will no longer be necessary, or humanly possible as we crank the agents up to 11.

But then how will we review each PR enough to have confidence in it?

How will we understand the overall codebase too after it gets much bigger?

Are there any better tools here other than just asking LLMs to summarize code, or flag risky code... any good "code reader" tools (like code editors but focused on this reading task)?

Most car manufacturers made this mistake because they started mimicking the then leader for innovation (and customer satisfaction), Tesla, too much.

General cautionary tale: just coz a company is successful, doesn't mean it's doing _everything_ right. Plenty of folks who love their Teslas would prefer a few more buttons (and door handles on the inside, etc) if given the choice. Could say similar things about some choices Apple made.

"Almost anyone can prompt an LLM to generate a thousand-line patch and submit it for code review. That’s no longer valuable. What’s valuable is contributing code that is proven to work."

I'd go further: what's valuable is code review. So review the AI agent's code yourself first, ensuring not only that it's proven to work, but also that it's good quality (across various dimensions but most importantly in maintainability in future). If you're already overwhelmed by that thousand-line patch, try to create a hundred-line patch that accomplishes the same task.

I expect code review tools to also rapidly change, as lines of code written per person dramatically increase. Any good new tools already?

Have you tried them with providing a grounding resource, e.g. attaching a file to ChatGPT or NotebookLM? Yes need some human expert to create (or curate) that grounding resource in the first place, but LLMs handle the rest well: presenting info in different ways and paces, interacting with the learner like a tutor, etc.

End of an era: video (with broadband Internet penetration) was the best tool we had for 15+ years. But LLMs are now good enough, including in image+infographic generation and factuality (especially when grounding resources are provided... which is where human experts still matter). I think video is now better only for learning physical hands-on skills... and those videos tend to be on YouTube rather than on Udemy or Coursera.

Coursera's model will still survive for a while, given people's desire for branded credentials (university degree credits or company-branded certificates)... until the university bubble bursts too in a 10+ years. Start of trend: https://www.nbcnews.com/politics/politics-news/poll-dramatic...

A bit of a plug: we tried building a consumer business, with a learning experience built atop these LLMs: https://uphop.ai/learn . Still offered for free to consumers, but we're now succeeding much better on B2B ("you either die a consumer business or live long enough to become B2B" was v true for us).

Glad to see big improvement in the SimpleQA Verified benchmark (28->69%), which is meant to measure factuality (built-in, i.e. without adding grounding resources). That's one benchmark where all models seemed to have low scores until recently. Can't wait to see a model go over 90%... then will be years till the competition is over number of 9s in such a factuality benchmark, but that'd be glorious.

Big knowledge cutoff jump from Sep 2024 to Aug 2025. How'd they pull that off for a small point release, which presumably hasn't done a fresh pre-training over the web?

Did they figure out how to do more incremental knowledge updates somehow? If yes that'd be a huge change to these releases going forward. I'd appreciate the freshness that comes with that (without having to rely on web search as a RAG tool, which isn't as deeply intelligent, as is game-able by SEO).

With Gemini 3, my only disappointment was 0 change in knowledge cutoff relative to 2.5's (Jan 2025).

It's all about the chip economics. I don't know how the _manufacturing cost_ of Google's TPUs compares to Nvidia's GPUs, for inference of equivalent token throughput.

But at the moment Nvidia's 75-80% gross margin is slowly killing its customers like OpenAI. Eventually Nvidia will drop its margins, because non-0 profit from OpenAI is better than the 0 it'll be if OpenAI doesn't survive. Will be interesting to see if, say, 1/3 the chip cost would make OpenAI gross margin profitable... numbers bandied in this thread of $20B revenue with $115B cost imply they need 1/6 the chip cost, but I doubt those numbers are right (hard to get accurate $ numbers for a private company for the benefit of us arm-chair commenters).

"complementing the Neural Accelerators in the CPU and GPU" seems to be a misprint; I don't believe they have the accelerators in the CPU too.

Still super interesting architecture with accelerators in each GPU core _and_ a dedicated neural engine. Any links to software documentation for how to leverage both together, or when to leverage one vs the other?

Plug for our https://uphop.ai/app : it's for adult learning / corporate training. We break down a desired job skill into small chunks, and engage the user with practice & give nuanced feedback. And of course like chatbots make it easy for user to ask more questions or go on tangents.

Would appreciate feedback!

There's a bit of overlap with Learn Your Way I guess. I'm not sure users need to toggle between alternate formats of the same instruction though. Instead the instruction itself should be as multi-modal as possible, and offer flexibility to ask questions... which even gemini.google.com offers so I'm not sure this is a net improvement over that.

Neural band is huge, glad they're shipping it already rather than waiting (years?) for a production version of Orion (the full AR glasses they demo'd a year ago together with this neural band). TheVerge found the controls great, even tried an alpha of handwriting for text input: https://youtu.be/5cVGKvl7Oek

These glasses are just "annotated reality" rather than full AR, with just 1 small display; think Google Glass but 100x more discreet. So discreet input and output on a device with a camera.

Neural band is huge, glad they're shipping it already rather than waiting (years?) for a production version of Orion (the full AR glasses they demo'd a year ago together with this neural band). TheVerge found the controls great, even tried an alpha of handwriting for text input: https://youtu.be/5cVGKvl7Oek

These glasses are just "annotated reality" rather than full Augmented Reality, with just 1 small display; think Google Glass but 100x more discreet.

I've thought larger bi-folds have an odd aspect ratio for anything but two-app multi-tasking. E.g. there's ~ no benefit for videos compared to non-foldables with half the screen area. Is aspect ratio not an issue in practice somehow?

Great points. Re safety, I wasn't imagining the consumer doing it themselves but some robot like ones imagined for whole-battery swapping (e.g. https://youtu.be/Oj6LaYFall4); I think any battery swapping only makes sense for batteries you rent rather than own.

Neat idea to mix batteries of different age and chemistries. I've wondered why EVs couldn't do that too with some power electronics and SW. If an EV battery could have multiple such modules, it'd:

1) Make it easier to carry a cheaper lighter less-natural-resources-consuming battery most of the time. Go to some "gas station" to rent and add more modules when taking a road trip

2) Make it cheaper to replace the 1 module used a lot at its EOL, thereby making EVs last longer and be viable as cheap used cars even past 10 years like ICE cars are

3) Allow easier upgrades as chemistry improves: solid-state, sodium ion, etc.

Modules could be electrically tested for fit. I'd think the fit range would be quite wide (e.g. if one supported lower max discharge rates than another) given the headroom we have with EVs' power these days: they have far-more-than-needed power (which mostly comes for free with EV range).

The tradeoff is that they'd need to be built to be modular with some standardization on module dimensions (maybe we'll have "ZZ" size like we have AA, C, etc today), and would take a tad more volume in the vehicle (though the limiting factor is weight rather than volume). Easily worthwhile over the current model with a huge monolithic pack.

Apple and Google are now competing on being more closed, rather than on being more open. Perhaps because we gave Apple a free pass on curbing our freedoms, and even defended its actions as needed for 'security'

AI interviews don't make sense when it's each company making the candidate do separate interviews, coz the company isn't spending any human time (no skin in the game) and can easily abuse candidates' time. I think it'd make sense tho if there were shared interviews, e.g. candidate does just one interview for all backend SWE roles (or all SWE roles that use AWS EC2 and Dynamo, or whatever further specialization)... any number of companies should then be able to access the AI insights of that interview, together with other candidates' interviews for that same role.

Then hiring manager can still bring the 2-5 best-fitting candidates onsite for 1-3 human-led interviews (hopefully fewer than what were needed before). The benefit of the AI interview would be to give way more signal than a resume can, making matching more efficient for both sides.

I had filled out the form for this. Wish Eric stuck to this instead of the Pebble revival: it'd have a bigger market.

I don't understand how the market isn't considered big enough for any phone OEM: how can it be smaller than that of foldables? Or even if it is, isn't it still big enough, and shouldn't there generally be more sizes and form factors of phones?

It's as tho the car industry decided to only make 184" long SUVs (6.2-6.7" phones) and 200" long 3-row SUVs (foldables)... no other SUVs, no sedans/hatchbacks, no sports cars (much smaller and much lower volume). And different cars are actually hard to engineer and mass-manufacture the chassis and bodies for... in contrast a phone's HW is inherently more modular and mostly just the screen and battery need to be changed for each size.