just a regular person trying every ai tool that drops and telling you what actually works
Joined July 2013
- Tweets294
- Following1.4K
- Followers685
- Likes264
Pinned Tweet
every week a new ai tool drops and everyone loses their mind. i test them like someone who actually has to use them at work and in real life.
👀 ux teardowns
🔍 the workflow step the demo always skips
✅ tools worth using after week one
💬 honest takes when it breaks.
MM6 Maison Margiela x Salomon has a second drop on Nov 17: the XT-BOOT, with a drawstring you can wear snug or loose, at ¥89,100 online only.
a boot with a drawstring is the most Salomon thing a fashion house could ask for.
Akamai launched Agent Sandbox in tech preview. it runs code an AI agent writes inside a locked box, with outbound network access blocked by default.
you choose which outside services and tools it can reach.
you still need to decide which connections your agent really needs.
forecasting usually means a GPU and a big model.
Mistral's Timestral comes in four sizes from 5M to 209M parameters and runs on a CPU, with a one day ahead forecast taking 10 ms on the smallest.
the missing step was a model small enough to run where the data already is.
Gap and Victoria Beckham dropped a 37 piece fall collection on October 23, with prices from $28 to $368.
knitwear, outerwear, denim and basics, so it is a wearable range, not a runway capsule.
a range with prices from $28 to $368 gives people a way in at every budget.
The missing step
an AI assistant inside the apps people already use is the easy part.
Apple is adding one to Keynote, Pages, Numbers and Logic Pro in beta, with MCP support on macOS 27.
the missing step was a way to connect it to everything else.
agents can browse, search and use web apps, but someone has to decide where they practice.
openai now trains frontier models with no internet access, on an offline copy of cached web content.
the missing step was a sandbox they cannot leave.
techcrunch.com/2026/10/23/op…
h&m x elie saab landed on october 22, menswear, womenswear and accessories.
sequins, lace and dégradé finishes, so red carpet looks without the red carpet price tag (prices not announced in the coverage i found).
would love to see how the menswear pieces look styled.
The World of Anna Sui opened Oct 21 at the Cranbrook Art Museum in Detroit with more than 100 looks from her archive, and it runs through Feb 28.
a designer known for mixing eras getting a full archive show feels right while brands everywhere are digging through their back catalogues. worth the trip if you are near Detroit.
tencent released WeVisOne-VL, one open model that handles six vision tasks, including detection, OCR, segmentation and matting. weights and code are on Hugging Face and GitHub.
in everyday terms: finding things in a photo, reading text, cutting out a subject and removing a background could all run through one model.
the missing step is checking how it runs on your own machine. look at how much memory it needs, and whether a model that does six jobs is fast enough to be worth switching to.
try it on ten of your own images before dropping the tools you use now.
aiweekly.co/alerts/tencent-w…
early 2000s denim is coming back from two directions.
Madhappy x Lucky Brand went live on Madhappy’s site on Oct 16 and lands on Lucky Brand’s site and app on Oct 23.
Acne Studios x Hussein Chalayan pulls from Chalayan’s early 2000s designs.
the trend goes past low rise jeans, brands are digging through their own archives.
mistral large 4 is in public preview at $1.36 per million input tokens and $4.18 per million output tokens. it is a mixture of experts model with about 1 trillion parameters, 52 billion active per request, and mistral says the weights follow by the end of October.
a million tokens is roughly 750,000 words, so casual use costs very little. the bill shows up when an agent runs all day.
the announcement does not name the licence, so check what you are allowed to do with the weights before building on them.
the benchmarks are mistral’s own, so test it on your own work first.
the Victoria’s Secret Fashion Show moved to Los Angeles for the first time on Oct 18, with an all women performer lineup of Megan Moroney, Megan Thee Stallion, Tate McRae and KATSEYE.
ten returning models, including Adriana Lima and Gigi Hadid, walked. the venue changed, the formula of music, models and a live stream stayed put.
your files stay on your Mac, while the agent can still actually get work done? . this is the privacy first AI experience we have been waiting!
openai says its frontier models and Codex workflows are coming to Google Cloud’s Model Garden in the coming weeks.
if you use either side, the useful question is what changes for you on day one.
check which models show up, whether Codex runs the same there as it does today, and how the billing appears on a Google Cloud invoice.
one catalog with models from both companies is easy to like. the product still has to show which model you picked and what it cost before you commit a project to it
aiweekly.co/alerts/openai-pu…
amazon’s new alexa tablets start at $229 and run full android with the google play store.
that matters most if you owned a fire tablet. you are no longer limited to amazon’s app store to find the tool you want.
before buying, check three things: which apps you actually use, whether alexa+ is included or needs a subscription, and how much storage the cheapest model has.
a cheap tablet with every app available is only useful if the one you need runs well on it
ai has made work easier for me, but only because it fits into the flow
AI was supposed to make us work less
I've never been this tired in my life
most days I have up to 10 agents running at the same time. one fixing a bug, one writing pages, one refactoring something I already forgot I asked for, 3 other I have no idea what they are doing but i suppose i ask it?
and I jump from conversation to conversation all day
compared to 5 years ago, I probably context switch 100x more
by the end of the day my brain is fried. completely toasted. nothing left, not even for the people I love
I call it the fried brain effect
thought it was just me but almost every founder I talk to lately says the exact same thing
we got superpowers, and somehow the pressure only went up. if you can ship 10x more, you feel like you have to ship 10x more. every single day
I don't have a solution
honestly I have no idea where this goes from here
anyone else feeling this?
apple’s HomeView identifies who approaches and changes the screen for them
clever idea, but shared devices need an obvious guest mode
if every useful screen depends on facial recognition, visitors and temporary users become a UX exception
personalisation still needs an escape hatch
