LLMs have lots of last-mile problems.
LLMs have lots of last-mile problems. Largely due to gardening the right context and the right coactive UI.
Topic: llms
1,749 mentions · 790 chunks · 131 episodes
184.9× distinctiveness vs baseline
?
How much more common this term is here than in ordinary English. Higher values mean the topic is more characteristic of this corpus.
1.5x burst in 2025 Q1
?
Peak quarter intensity across the topic's active span. Higher values mean attention was concentrated into a shorter stretch rather than spread evenly over time.
Topic summary
?
A short read on the topic's time range, peak episode, and strongest associations. Use it as the quick orientation before drilling into examples.
- llms appears in 790 chunks across 131 episodes, from 2023-11-06 to 2026-08-03.
- Its densest episode is Bits and Bobs 10/20/25 (2025-10-20), with 17 observations on this topic.
- Semantically it travels with ChatGPT, Claude, and Google, while by chunk count it sits between Claude; its yearly rank moved from #2 in 2023 to #1 in 2026.
Over time
?
Raw mentions over time. Use this to see absolute attention, not relative rank among all topics.
Observations
?
The primary evidence view for this topic. Sort it chronologically when you want concrete examples behind the larger pattern.
If you've gone through the effort of having high quality programmatic thinking LLMs can write infinite Op Eds for you, on demand.
If you've gone through the effort of having high quality programmatic thinking LLMs can write infinite Op Eds for you, on demand. The backlog of Bits and Bobs feels exceptionally valuable to me in the age of AI.
Bruce Schneier points out that LLMs will bring mass spying.
Bruce Schneier points out that LLMs will bring mass spying. Before, we had mass surveillance, but a human sifting through the collected data happened rarely. That limited the oversight ...
LLMs are infinitely patient so good enough ACLs aren't good enough any more.
LLMs are infinitely patient so good enough ACLs aren't good enough any more. Before your data was protected a bit by security through obscurity. But LLMs ...
Garry Tan on Twitter:
Garry Tan on Twitter: "New social networks are going to appear that will be LLMs creating a cozy web customized for us and our real friends, and their friends and so on There will be a new social network built on mutual trust, all...
One of the downsides of infinite patience of LLMs: encouraging spirals in people with OCD.
One of the downsides of infinite patience of LLMs: encouraging spirals in people with OCD. Sycosocial.
LLMs are like tofu.
LLMs are like tofu.[gj] They absorb the intention of the context. Even an earnest and helpful LLM, when given malicious context, can itself become malicio...
Which will be more important by unit weight in software systems in the AI era, LLMs or normal code?
Which will be more important by unit weight in software systems in the AI era, LLMs or normal code? A lot of platforms being built for the age of AI imagine that most of the weight of systems will be LLMs, with just a little bit of c...
Which is easier, engineers to become subject matter experts, or subject matter experts becoming engineers?
...tem pole and always will be". But the tech industry as a whole would miss it if LLMs changed it because we'd all have the same bias and blindspot. Domains that don't work with CS are less deterministic. Their taste, metacognition, ent...
A lot of the best practices for programming with LLMs are the same as for humans.
A lot of the best practices for programming with LLMs are the same as for humans. For example, document each directory with a README, have aggressive type checking / linters. But most humans give up or a...
Computer Science was used to build LLMs.
Computer Science was used to build LLMs. But Computer Science can't tell us how to understand LLMs.
LLMs are essence extractors, not mechanical reproducers.
LLMs are essence extractors, not mechanical reproducers. The way they learn from their training data is more than just reproducing. Essence is a new conce...
I love Ben Follington's piece on interacting with LLMs as dream walking.
I love Ben Follington's piece on interacting with LLMs as dream walking.
Every built a Diplomacy game for LLMs to play.
Every built a Diplomacy game for LLMs to play. Gemini generally does well. Claude refuses to lie, and thus loses often. ChatGPT o3 often wins because it is very happy to betray its collab...
You can typically trust off-the-shelf LLMs to not try to manipulate you in particular.
You can typically trust off-the-shelf LLMs to not try to manipulate you in particular. But LLMs are easy to fool. So if anyone else you don't trust is feeding input into the context, then the ...
I used ChatGPt to identify an odd piece of hardware in my post-war AirBnB in Manhattan this week.
...hatGPT. Only an open-ended tool could do that. Now that we have approaches like LLMs it's impossible to see how Google Lens's old approach of manually curating data for individual verticals could never have possibly ever worked in gen...
For intentional tech, you can get far with a short essay that distills your values.
...ge people would care about or write. Now how are you the same as everyone else (LLMs already have that baked in) but how are you different?
I think LLMs will likely make corporate politics worse.
I think LLMs will likely make corporate politics worse. The metagame will just get more inscrutable and energized. Corporate politics emerge from the fundamental ...
A corporate politics maneuver classically used with consultants will also be applied to LLMs.
...ate politics maneuver classically used with consultants will also be applied to LLMs. You bring in a consultant to come to the conclusion you secretly believe. If your boss agrees with the consultant's conclusion, then you get the ben...
A pattern: use LLMs in tiny tasks where the overall swarm is emergently powerful.
A pattern: use LLMs in tiny tasks where the overall swarm is emergently powerful. LLMs handle small tasks very well. The emergent swarm might lose the plot (e.g. forgett...