vertis

12 ms·

91.
by vertis 3y ago
Llama2:7b makes the same mistakes. It's not until you use something like Mixtral or Llama2:13b that it actually gets the correct results (in my one example). Interestingly Llama2:13b objects that there is no discount until I clarify: &
92.
by vertis 3y ago
It might be equivalent to LLaMA2 but it's still not capable of even simple reasoning: > If two apples cost 2 dollars and 1 apple costs 1.20, what is the discount you're getting > To calculate the discount, you need to compar
93.
by vertis 3y ago
I disagree, having used Sweep extensively, I've found the GitHub Issue -> PR flow to be incredibly clunky with a lack of ability to see what is happening and what has gone wrong.
94.
by vertis 3y ago
What you want here is a local uncensored model. Preferrably one you've trained from scratch otherwise a government could have put in bad information that would cause your revolutionary efforts to fail.
95.
by vertis 3y ago
People (SWEs) don't want to hear this. I think it's an inevitability that something of this nature will happen.
96.
by vertis 3y ago
Conservatively, I think LLMs will replace SWE roles within the next 5-10 years. In that period SWE roles will change drastically, to be more hearding of AI agents. We can't hope to compete with something that can edit all the files in
97.
by vertis 3y ago
I fell down reading about all the ships that have hit bridges, and this is probably the most practical solution for a lot of scenarios.
98.
by vertis 3y ago
This seems fairly focused on safety, and less on the working in the title. But working from Hotels is definitely a pain. They almost always have worse wifi than an AirBnB. It varies whether they have a good seat, table that you can work at
99.
Reflections on a Startup Misstep: A Decade of Perspective(vertis.io)
1 points by vertis 3y ago|0 comments
100.
by vertis 3y ago
I've had some very close relationships with C levels at some big and medium companies. I've caught them lying in varying degrees more than once. Sometimes though the job demands it. They're potentially not allowed to disclose
101.
by vertis 3y ago
Right, that definitely makes it much harder.
102.
by vertis 3y ago
I'm curious what subset of tech you're applying for, and what your region and salary expectation are. Speaking as someone with 20+ YOE experience myself.
103.
by vertis 3y ago
Why do you want to be a <insert job you're passionate about>, just do it as a hobby spend ~8 hours a day doing something you hate.
104.
by vertis 3y ago
When the hobbies become indistinguishable from the job. Congrats on the reentry.
105.
by vertis 3y ago
There is only so much you can do though if the other people aren't prepared to learn how to use a password safe (etc).
106.
by vertis 3y ago
It's probably some combination of both. There is a subset that can be reformed and made bad decisions, and there is a subset that are beyond help and just need to be kept from further damaging society. (fwiw: I'm very much in the
107.
by vertis 3y ago
There is a future where they're locked out of the servers.
108.
by vertis 3y ago
I use Devyce[0] which has a UK phone number. Works for all the 2FA codes I've ever needed. Can port your existing UK number as well. They've been my provider for 3 or so years now without any problems. Runs fine on top of Airalo w
109.
by vertis 3y ago
Admittedly, I'm not a huge user of mobile web (I'm not an iPhone user either, I'm still using a Pixel 4a). However, I am absolutely not going to defend extremely heavy javascript content either. There needs to be a balance, a
110.
by vertis 3y ago
I think that's the point, they released GPT2 openly, but as soon as they had something commercially viable they became ClosedAI.
111.
by vertis 3y ago
It's really easy to tell the difference between a democracy and a fake democracy. Democracies are messy, people never agree. Anywhere that get's consistent landslides for one person or party is not a democracy. Take for example Fr
112.
by vertis 3y ago
Which is how we end up with another framework :)
113.
by vertis 3y ago
This article makes incredibly poor arguments to support the title. "destroyed the web" is such a massive stretch - I remember pre-javascript internet and I remember pre-css internet, coded my first HTML page in 1995. Even early J
114.
by vertis 3y ago
So many flaws with this article, hard to know where to start. So many unsubstantiated arguments. It jumps from React devs not knowing meta tags and links, to oddly hyping WordPress and trying to sideline actual solutions like Next and Svelt
115.
by vertis 3y ago
I stopped using Spotify a long long time ago, for much more selfish reasons. I would try and "download" a bunch of songs before an international flight only to hop on the flight and discover that despite saying they were downloade
116.
by vertis 3y ago
Glass half empty then. Most vector databases are using some local or external provider to get the embeddings and then using some storage engine to store and retreive the embeddings. Whether it's pgvector leaning on postgresql, or chrom
117.
by vertis 3y ago
That's a slippery slope argument, which is a common logical fallacy[0]. This model being inefficient compared to the best possible implementation does not mean that future additions will also be inefficient. It's the equivalent to
118.
by vertis 3y ago
I mean if you care about that you shouldn't be running anything that isn't highly optimized. Don't open webpages that might be CPU or GPU intensive. Don't run Electron apps, or really anything that isn't built in a
119.
by vertis 3y ago
No, I wrote all this myself...
120.
by vertis 3y ago
And you're biased in thinking everything has bad public transport like the US. I could use an absolute bucket load of examples from UK, Europe and Australia where to drive long distance you will always come off second to public transpo

More ›