AI Gets Practical, Chips Go Wearable

I've been working with local models since they came out, and finally, they're surprisingly good now.

Early on, models were slow, hard to use, and just not that accurate for most programming tasks. The idea that local models were severely lagging behind was largely true until, for me, the release of GPT-OSS.

As a result, I've mostly been using local models as fast, personalized Google for development questions that don't require recency.

But with the most recent releases from Google in the Gemma 4 , family, I've finally been able to do agentic coding locally and have loops work at about ~75% the accuracy/speed of frontier models, which is incredible.

I've so far been using gemma-4-26b-a4b LM Studio implementation as my default local model. I've used the local setup so far to: Refactor a Python script that was a notebook into a repo of 5-6 modules, lint that module to use correct type hints for generics (most frontier models now do this automatically, but not always).

Artificial Analysis Artificial Analysis Models Coding Agents Speech, Image, Video, Music Hardware Leaderboards About AI Trends Arenas K All articles June 17, 2026

Z ai’s GLM-5.2 is the new leading open weights model on the Artificial Analysis Intelligence Index scoring 51 and it sits on the Pareto frontier of Intelligence vs Cost per Task

GLM-5.2 is the same size as GLM-5.1 (744B total / 40B active parameters) but scores 11 points higher on the Intelligence Index v4.1, placing ahead of MiniMax-M3 (44) and DeepSeek V4 Pro (max, 44). On the first-party API it is priced in line with GLM-5.1 at $1.4/$4.4/$0.26 per 1M input/output/cache hit tokens

➤ GLM-5.2 is the leading open weights model on the Intelligence Index v4.1. At 51, it leads MiniMax-M3 (44), DeepSeek V4 Pro (max, 44) and Kimi K2.6 (43)

➤ Improvements across most evaluations, particularly scientific reasoning: GLM-5.2 gains over GLM-5.1 on most evaluations, led by scientific reasoning on CritPt (+16 points to 21%) and HLE (+12 points to 40%), alongside AA-LCR (+9 points to 71%), tau3 banking (+15 points to 27%) and SciCode (+7 points to 50%). TerminalBench v2.1 also improves (+16 points to 78%) and GPQA Diamond gains 3 points to 89%

Qualcomm CEO Cristiano Amon said Tuesday that the company is working on over 40 different AI wearable devices — including jewelry, earbuds with cameras, pins, and watches — a sign of how aggressively the chipmaker is betting that the next major computing platform won’t be a phone.

To power that vision, Qualcomm is announcing two new offerings: a platform called Snapdragon Reality Elite for mixed-reality glasses, designed to run more powerful on-device AI, and the Scalable Turnkey AI-Ready Toolkit (START), a combination of hardware modules and a software stack for AI devices, starting with smart glasses.

Compared to its previous XR platform, the new Snapdragon Reality Elite delivers improvements of up to 60% in GPU performance, up to 30% in CPU performance, and up to 160% in NPU performance, according to the company.

The Snapdragon Reality Elite supports 4.4K per-eye resolution at 90 fps, a modest bump from the XR2+ Gen 2’s 4.3K per-eye resolution. (The higher the per-eye resolution and frame rate, the sharper and smoother the visual experience, which matters most for reducing the motion sickness and eye strain that’ve historically made extended headset use uncomfortable.)

Qualcomm says the platform is designed to power two types of devices: stand-alone video-see-through (VST) headsets, which layer digital content over a camera feed of the real world, and lightweight, tethered optical-see-through (OST) glasses, which blend digital imagery directly into your field of view.

If you've got a bit of time to watch a presentation on it, I highly recommend this talk: www.youtube.com/watch?v=pYeekwv3vC4 (Note that other topics are largely skimmed over, such as CSRF protection. You should learn about other topics from other sources. Also note that "valid" usecases for JWTs at the end of the video can also be easily handled by other, better, and more secure tools. Specifically, PASETO .)

A related topic: Don't use localStorage (or sessionStorage) for authentication credentials, including JWT tokens: www.rdegges.com/2018/please-stop-using-local-storage

But stateless is better! You can't securely have truly stateless authentication without having massive resources, see the cryto.net link above. Also, Stateless is a lie .

If you do need a short-lived, signed token for something, there is a better spec called PASETO which is designed to be secure. Just make sure you aren't using them for sessions.

Copy link Copy Markdown Author "Sessions don't make any sense in a stateless API."

Uber plans to launch a premium robotaxi service in Houston by mid-2027, making it the second U.S. market under its partnership with EV maker Lucid and autonomous vehicle startup Nuro.

The announcement follows a flurry of activity in the San Francisco Bay Area, as the trio of companies prepare to offer a robotaxi service there later this year. Uber says it will eventually take the robotaxi program to “dozens of cities” in the coming years.

For now, the focus is on San Francisco and then Houston — both markets where Uber will go head-to-head with rival Waymo, the Alphabet-owned autonomous vehicle company that currently operates commercial robotaxi services in both cities.

Nuro has spent months testing Lucid Gravity SUVs equipped with its self-driving system in San Francisco and has made progress on that front, including giving Uber employees the ability to hail the Lucid robotaxis. But these vehicles still aren’t driverless, despite Nuro receiving a permit last month from the California Department of Motor Vehicles that would allow it to remove the safety driver from the vehicle.

Uber and Nuro’s combined engineering fleet of 100 autonomous vehicles is testing on public roads with safety operators behind the wheel in Houston as well. Nuro is also using closed courses and simulation to validate the self-driving system before opening the robotaxis to the public.