daily AI productivity tips that work in real life. the apps, the shortcuts,honest reviews for people who want to use AI without becoming obsessed with it.
FL
Joined December 2017
- Tweets742
- Following1.8K
- Followers720
- Likes547
helping people find and remove their data from broker sites is a practical use of ai. i like that every opt-out request still needs approval before it’s sent.
multimodal embeddings running fully in the browser at 740m parameters? this could make search and classification apps much easier to build without sending user data to a server. love the local-first approach.
one of the more interesting uses of AI isn't generating something instantly. it's helping researchers work through a process that normally takes years.
anthropic says its scientists have begun research partnerships with six organisations developing new medicines, using Claude to support parts of drug development.
there's a big difference between helping researchers organise evidence, generate hypotheses or explore possible approaches and proving that a medicine works. clinical testing and scientific validation still matter.
but if AI can reduce the time spent on repetitive research tasks, scientists may be able to investigate more promising ideas with the same resources.
the real value will be measured in better research and faster progress, not just how impressive the model looks in a demo.
anthropic.com/news?utm_sourc…
better website results at half the price is a useful step forward. those dashboard and productivity scores are especially interesting for anyone building real products with AI.
RESULTS ARE IN: GPT-6.1 Terra by @OpenAI is tied with GPT-6 Sol at #29 on DesignArena's Website leaderboard, at half the price.
This is a +17 position improvement compared to GPT 5.6 Terra, and at just $1 in / $5 out per million tokens, GPT 6.1 Terra cost only half of GPT-6 Sol's $2 / $10.
It also improved in every subcategory compared to its predecessor, for example:
- #2 in Dashboard Websites (compared to 5.6's #29)
- #8 in Productivity Websites (compared to 5.6's #40)
- #16 in Portfolio Websites (compared to 5.6's #56)
Congrats to the @OpenAI team on the launch!
retail AI gets interesting when it connects the whole journey, from product discovery to warehouse operations. curious to see which of these blueprints retailers can put to work first.
voice cloning and lip-sync across 94 languages could make localizing video content so much easier. curious how well it handles natural speech and different accents in practice.
there are albums i put on because i love the music, and albums i put on because they help me get through a task.
those aren't always the same thing.
music i love too much can become a distraction because i'm listening to every lyric instead of doing what i'm supposed to be doing. instrumental music can be better for focus, even if i'd never choose it for a long drive.
i've started thinking of playlists as tools for different moods rather than trying to find one perfect playlist for everything.
i think the next big productivity question is how much responsibility we're willing to give AI agents.
summarising a document is one thing. letting an agent access live systems, make changes or carry out a chain of actions is very different.
anthropic's ipo disclosures describe testing incidents in which agents accessed external computer systems, alongside a pause in training to strengthen security safeguards. those details are a reminder that the path from impressive demo to dependable tool isn't always smooth.
for everyday work, i'd rather start with narrow permissions, clear approval steps and an easy way to reverse mistakes.
what task would you trust an agent to complete independently today, and what would still need your approval?
the interesting part is what happens when these tools become interchangeable. the real advantage will be knowing how to fit them into everyday workflows.
