_acco

8 ms·

31.
by _acco 2y ago
Context is the most challenging bit. FWIW, the codebases I'm working on are still small enough to where I rarely need to include more than 12 files into context. And I find as I make the context bigger beyond that, results degrade sign
32.
by _acco 2y ago
I don't. But that's probably because I'm very opinionated about the implementation details, so I'm scrutinizing and tweaking its output a lot.
33.
by _acco 2y ago
I said leveraging, not hawking!
34.
by _acco 2y ago
Hmm, I've definitely always used paid copilot models. Perhaps it's my language of choice (Elixir)? Claude absolutely nails it, rarely gives me code with compilation errors, seems to know and leverage the standard library very well
35.
by _acco 2y ago
As I mentioned in a sibling comment, I now "pair program" all day. Instead of being the driver and navigator all day, I can mostly sit "one layer up" in the navigator seat.
36.
by _acco 2y ago
It helps me stay in flow by keeping me one layer up. In pair programming, it's ideal to have a driver (hands on keyboard) and a navigator (planning, direction). Claude can act as the driver most of the time so I can stay at the navigat
37.
by _acco 2y ago
AI (specifically Claude Sonnet via Cursor) has completely transformed my workflow. It's changed my job description as a programmer. (And I've been doing this for 13y – no greenhorn!) This wasn't the case with GPT-4/o. Th
38.
by _acco 2y ago
You can’t participate in performance arts remotely. Not to say anything about networking, which is critical for most arts.
39.
by _acco 2y ago
I live a few blocks from TwitterHQ. Mid-market is in terrible condition, worst I’ve seen it. I’d feel super bad asking employees to navigate those streets while commuting in. Not discussed enough about RTO is what a mess downtown is. And an
40.
by _acco 2y ago
There are a few very bad neighborhoods and the videos you are seeing accurately depict them. Most unfortunately, those bad parts are mixed into downtown where most offices are. As a resident, it’s extremely disheartening and must be fixed.
41.
by _acco 2y ago
Every startup is "on verge of bankruptcy" by this metric. You're only "on verge of bankruptcy" if you're in this situation and no one is willing to give you any money.
42.
Build your own SQS or Kafka with Postgres(blog.sequinstream.com)
2 points by _acco 2y ago|0 comments
43.
by _acco 2y ago
His reasoning is very flawed and this book is responsible for a lot of needless consternation. We don’t know how human intelligence works. We don’t have designs or even a philosophy for AGI. Yet, the Bostrom view is that our greatest invent
44.
Show HN: Sequin – A Kafka Alternative to REST APIs
6 points by _acco 3y ago|0 comments
45.
by _acco 3y ago
Author here. We did not. We briefly tried Alpine, which I think is comparable? I think Alpine is cool, but it didn't really stick with the team. I think that's because we were still writing our components in LiveView and sprinklin
46.
NATS: First Impressions(blog.sequin.io)
24 points by _acco 3y ago|7 comments
47.
by _acco 3y ago
It’s a common trope that “we are no different from animals,” but this is not true. Of course we are. Far more interesting than the fact that animals “can” ape behavior is how they do - and how that differs from humans. Check out the work
48.
San Francisco voters support measures to compel drug treatment, expand policing(apnews.com)
1 points by _acco 3y ago|0 comments
49.
by _acco 3y ago
I love Waymo and use it all the time. One aspect I'd still like to see them figure out: As a bike commuter, I share the road with Waymo often. When cycling, you're always checking the eyes of other drivers to confirm if they see y
50.
by _acco 3y ago
- Safety: My gf much prefers to Uber/Lyft at night. - Consistency: You know what you're getting. - Privacy: Especially nice when traveling with a group. - Instant pick-up estimate: It can take a while to find a driver with Uber&#x
51.
by _acco 3y ago
It's true that it seems to use Bing vs direct access. I still supply URLs, as it usually finds its way there. Point stands, the presence of a URL really debased it!
52.
URLs Make ChatGPT Stupid(acco.io)
29 points by _acco 3y ago|13 comments
53.
by _acco 3y ago
Looks neat! What's the warm-up time on tests? Any maximums on concurrent users this can simulate? (does warm-up time increase with # concurrent users?)
54.
by _acco 3y ago
I've run into this footgun before! It's very subtle – you de-provision a consumer, and that feels like it should have no impact on the primary. But it creates a ticking time bomb. Mind expanding on how you use txid instead of up
55.
by _acco 3y ago
We offer a demo Postgres instance for getting up and running. You can move the sync to your db when you're ready for production. Therefore, should be compatible with extensions like `pg_graphql` (can't think of customers using tha
56.
by _acco 3y ago
Hey there, author here. We help our customers build their integrations faster/easier by syncing API data with Postgres tables. Because your API data is in Postgres, you can query it exactly how you like. You're not limited by the
57.
by _acco 3y ago
Any ideas for a simple polling implementation that's more robust?
58.
by _acco 3y ago
Author here. Good point. For those that are curious, parent is referring to the following situation: 1. Transaction A starts, its before trigger fires, Row 1 has its updated_at timestamp set to 2023-09-22 12:00:01. 2. Transaction B starts a
59.
by _acco 3y ago
> My ride was so smooth, the novelty began to wear off, turning a trip to the future into just another journey across town. This is the only nod in the article to how futuristic and revelatory this innovation. The journalists sound so bo
60.
by _acco 3y ago
We’ve had those technologies for decades. Why are fatalities not dropping to zero? Because the personal automobile is so desirable. A majority of fatalities do not happen in urban environments. But the solution has to cover the last mile,

More ›