jasfi

8 ms·

61.
by jasfi 2y ago
I like the style of the game, minimalist with a focus on language. It would be great to see this style with a story-line that doesn't intend to irritate the player.
62.
by jasfi 2y ago
I find Nim to have excellent trade-offs for writing system software. It's so much easier to write than Rust, but is also safe.
63.
by jasfi 2y ago
The pics help to disambiguate some of the titles, but its nowhere near as lean as HN.
64.
by jasfi 2y ago
I'm working on an AI agents platform that intends to reduce the amount of code you need to write to get high performing prompts working correctly. The wait-list is at https://aiconstrux.com
65.
by jasfi 2y ago
It's only called "raw" when called from an ORM, as opposed to generated SQL.
66.
by jasfi 2y ago
This is cool, but most ORMs have support for raw SQL.
67.
by jasfi 2y ago
I had a similar reply to someone else: by the time we can adapt to live on Mars, decades away at best, traveling there will be the least of our worries. Glad to be proven wrong, of course. Maybe we can try and terraform it in the meantime.
68.
by jasfi 2y ago
If we can pull off synthetic bodies for our consciousness, then by that time going to Mars will be a walk in the park. But I think you underlined the issue of timelines. Yes we can technically go to Mars in only a few years, but people livi
69.
by jasfi 2y ago
Just because Mars is better than the others doesn't mean its good enough as a long term habitat. People would have bone density problems much sooner in life, as an example.
70.
by jasfi 2y ago
The objective would likely drive lots of improvements to science and engineering. Mars itself is very inhospitable to humans.
71.
by jasfi 2y ago
The same thing will happen with this app. The user will select any answer to just do what they wanted to do.
72.
by jasfi 2y ago
Things that are difficult to do in tech, e.g. AI/ML that's in-house (not using an API).
73.
by jasfi 2y ago
Not too broken, or you waste a lot of time fixing bugs. Rather build prototypes minimally but working well.
74.
by jasfi 2y ago
New products or features are of an acceptable performance only, sometimes not even that. Additional work on performance tuning is often followed up in a later version. It's really about priorities that have to be decided, and deadlines
75.
by jasfi 2y ago
I'll probably take a chance when I'm ready.
76.
by jasfi 2y ago
I'm not OP, but I'm working on an AI powered NoCode platform. Could I contact you once I have a prototype going?
77.
by jasfi 2y ago
I'm competing in this space, though my solution is still in development. There's a wait-list you can sign-up for if you like: https://aiconstrux.com
78.
by jasfi 2y ago
The message is so obviously meant to insult people from an AI, that I suspect someone found a way to plant it in the training material. Perhaps some kind of attack on LLMs.
79.
by jasfi 2y ago
A lot of the top researchers are working on making LLMs more capable, so it's not impossible for new breakthroughs to occur, they might not just be as rapid paced as the last two years or so. There's also lots of utility to be fou
80.
by jasfi 2y ago
Google TPUs should theoretically be able to compete with Nvidia's H100, but for some reason they aren't seen as a practical alternative. I agree that training LLMs will get cheaper, but it's likely that the compute bottleneck
81.
by jasfi 2y ago
You can think of each of those as a bottleneck. The architecture (LLMs, transformers) was once the bottleneck, as was the amount of compute. From what I know the new bottleneck is the amount of quality data. Actually there was a breakthroug
82.
by jasfi 2y ago
Assisting with breakthroughs isn't sci-fi, it's things like bouncing ideas off an AI or having it help with code. It can do this to a limited degree. I did mean "help with" in a very practical, limited way.
83.
by jasfi 2y ago
That's an agent, which makes use of an LLM. The LLM's architecture means its trained once-off.
84.
by jasfi 2y ago
They likely have a high level plan to get to AGI, since this would help to give their research direction. I suspect stage 1 is to get current models to assist with making AI breakthroughs.
85.
by jasfi 2y ago
If you write a design doc first, you can save yourself a lot of time. It's very quick to write a new design, and designs are also much faster to revise than code is.
86.
by jasfi 2y ago
Maybe they don't feel comfortable calling it GPT-5 if it's not much of an improvement over 4o.
87.
by jasfi 2y ago
Very likely the training data.
88.
by jasfi 2y ago
My view is that LLM performance looks like bursts of improvement as breakthroughs happen, with lots of much smaller incremental improvements along the way. What OpenAI were hoping for was AI stepping in to provide the breakthroughs. But for
89.
by jasfi 2y ago
This is the sort of thing that LLMs should excel at. An AI coding assistant should theoretically help you move through these issues very quickly.
90.
by jasfi 2y ago
There's not enough space to represent code graphically, anything significant becomes crowded and messy. Every project like this starts off extolling that we are visual creatures, but this is a type of visualization we're not well

More ›