roseway4

A rambling screed. The article does include some interesting historical nuggets, but gets too hysterical for my taste when discussing present Silicon Valley.

PurpleAir?

Fascinating that a large portion of the Indian Ocean surface is ~100m lower than the global “sea-level”.

The project is clearly licensed under the BSD License. See the LICENSE file and the in-file comments below. You can't both license under the standard BSD License and carve out commercial usage.

""" authlib ~~~~~~~

    The ultimate Python library in building OAuth 1.0, OAuth 2.0 and OpenID
    Connect clients and providers. It covers from low level specification
    implementation to high level framework integrations.

    :copyright: (c) 2017 by Hsiaoming Yang.
    :license: BSD, see LICENSE for more details.

"""

Softbank had a strategy of deploying a lot of capital very quickly. That is, taking many more bets than traditional “high conviction” VCs. High transaction costs (including time to close) as a consequence of deep diligence would have broken this model.

It doesn’t look like this strategy worked out well for them.

I agree that some element of social housing is needed. Do you have any evidence to the assertion that this housing would not be respected by Americans?

Study referred to in the article: https://homelessness.ucsf.edu/our-impact/our-studies/califor...

> High housing costs and low income left participants vulnerable to homelessness. In the six months prior to homelessness, the median monthly household income was $960. A high pro- portion had been rent burdened. Approximately one in five participants (19%) entered homelessness from an institution (such as a prison or prolonged jail stay); 49% from a housing situation in which participants didn’t have their name on a lease or mortgage (non-leaseholder), and 32% from a housing situation where they had their name on a lease or mortgage (leaseholder).

tl;dr Nimby housing policies (and regressive property taxation) have driven this crisis.

tl;dr The prospect of paying significantly higher mortgage rates is dampening homeowner interest in selling. Despite these rates, demand is sufficient to be keeping prices high.

OpenAI just reduced its embedding API price significantly. It's a go-to for many developers, but it's not always the right choice. Slow response times can really erode user experience. In this analysis, I compared latency across popular APIs and models.

Great question. We have a deletion endpoint that allows sessions (think users) to be deleted. All memory artifacts are timestamped, and we run soft-deletes with deleted_at timestamps. This allows enterprises to execute retention strategies (keep all records for X months) and hard deletes / expungement at scale set by policy (all soft deletes must be expunged within 48 hours of a request).

One of the benefits of using Postgres + pgvector, is it is much easier to delete records from a vector index versus using a vector database or Redis. While Redis has ttls for objects, there isn't an easy way to remove data from a vector index.

Logically at the same time in the sql query. This may not be how the query is executed by the database, though, and is up to the optimizer.

We embed Postgres and pgvector in Zep[0], a memory store for LLM apps. It’s rock solid, even with very wide vectors, and with the right indexes ANN is performant up to tens of millions of rows. Haven’t tested beyond this, though.

It’s super useful for hybrid search: vector similarity search over unstructured data joined with filtering by structured data.

One caveat: It’s important indexes are added after a fair number of vectors have been persisted, which can make some production use cases challenging to implement.

[0] https://github.com/getzep/zep

It’s an advertising term. Unsurprising, given the context.

https://www.investopedia.com/terms/f/flighting.asp

Anybody know why Chromebooks would have <4 year OS support?

> The museum also promised a total rehang of its collection, as well as more “phygital” experiences (offered in physical or virtual format) and a refurbished library.

“phygital” is one of the more ridiculous conjugated neologisms I’ve come across.

Many of the "long-term memory" services focus on vector search over a corpus of documents to offer the LLM domain context. Where Zep is different, is we focus on the conversation history, offering the stateless LLM a long-term state component (alongside enriching the message history with summaries, metadata, token counts etc).

Re (2): GPT4 requires engineering for resiliency. The API (currently) has availability issues and high variance in latency. It’s not (yet) ready for interactive use. For other use cases, a good queuing and retry strategy is necessary.

Is it any better or worse than delegating work to a junior team member and not bothering to review the work product?

I love the idea that I can rapidly produce a rough draft of a document from notes, something I’m already doing with ChatGPT, albeit not as elegantly as if the functionality were integrated into Docs/Word.

After the second, closer, approach of the Russian fighter, a blade on the drone’s propeller is badly bent.

Yes, in the case of large retail banks, the government did help. In 2008, the federal government purchased stock in several banks in order to recapitalize them. One could argue this was a starker example of moral hazard than letting the bank fail, but making depositors whole.

It’s possible to do this either by fine-tuning an existing model or using an existing chat model prompts enriched by a vector search for relevant code. See my comment elsewhere.

OpenAI Codex understands code. Though it’s primary use case is code completion, it might be to do Q&A well given a prompt with context.

https://platform.openai.com/docs/guides/code

I’d you’re interested in trying the very cheap models behind ChatGPT, you may want to have a look at langchain and langchain-chat for an example of how to build a chatbot that uses vectorized source code to build context-aware prompts.

Rumors that sharks may be circling:

https://techcrunch.com/2023/03/09/silicon-valley-bank-shoots...

> Dish said its patents covered adaptive bitrate streaming technology that lets users stream content from around the world in real time "at the highest possible quality".

Appears this may be the patent: https://patents.google.com/patent/US9407564B2/en

Regarding the hard way, this little utility does a great job of splitting larger than memory JSON documents into collections of NDJSON files:

https://github.com/dolthub/jsplit

> ... last year Amazon had $38bn of ad revenue, more than any traditional media company and bigger than the entire (now much reduced) global newspaper industry.

> $38bn is a small percentage of Amazon’s overall $502bn 2022 revenue, but it probably had well over 50% operating margins, which would mean it brought in as much operating profit as AWS - $20-25bn, and without anything like as much capex.

Fascinating analysis. Ben also makes an interesting argument that Amazon uses ads as a price discrimination tool, allowing the store to extract more wealth from brands willing and able to pay for ads vs other sellers.